23- Statistics & Medicologeal Aspects Explains Flashcards
1
Q
What is clinical audit?
A
Clinical audit is a quality improvement process that aims to improve patient care and outcomes. It involves systematically reviewing care against explicit criteria and implementing changes as needed. It evaluates aspects of care such as structure, processes, and outcomes.
2
Q
What is the purpose of clinical audit?
A
The purpose of clinical audit is to improve healthcare delivery by identifying areas for improvement and implementing changes to achieve better patient care and outcomes.
3
Q
How is clinical audit conducted?
A
Clinical audit involves selecting specific aspects of care to evaluate against explicit criteria. Changes are implemented at the individual, team, or service level, and further monitoring is used to confirm improvement in healthcare delivery.
4
Q
What is research?
A
Research aims to derive new knowledge that has the potential to be generalizable or transferable. It involves conducting systematic investigations and studies to explore and uncover new information or insights.
4
Q
What is the goal of research?
A
The goal of research is to contribute to the body of knowledge in a particular field by discovering new information, generating evidence, and advancing understanding. It aims to expand knowledge and potentially improve practices or outcomes in various areas.
5
Q
What is a financial audit?
A
A financial audit is a historically oriented evaluation conducted to attest to the fairness, accuracy, and reliability of financial data. It is independent and aims to provide assurance on the financial information of an organization.
6
Q
What is an operational audit?
A
An operational audit is a future-oriented evaluation of organizational activities. It is systematic and independent, focusing on assessing operational policies and achievements related to organizational objectives. While financial data may be used, the primary sources of evidence are the operational aspects of the organization. This type of audit may evaluate internal controls and efficiencies.
7
Q
What is a departmental review?
A
A departmental review is an analysis of administrative functions during the current period. It aims to evaluate the adequacy of controls, safeguarding of assets, efficient use of resources, compliance with laws and regulations, and the integrity of financial information.
8
Q
What is a standards-based audit?
A
A standards-based audit involves comparing care or the passage of care against predefined and widely agreed standards or outcomes. It assesses whether the care provided meets the established standards.
9
Q
What is a systems-based audit?
A
A systems-based audit evaluates the processes occurring within an institution. It is an integral part of the clinical governance process and focuses on assessing and improving the systems and processes in place within an organization to ensure quality and safety of care.
10
Q
What is a Type 1 error in statistics?
A
A Type 1 error occurs when a statistical test incorrectly rejects a true null hypothesis. In other words, it is a false positive result where the test concludes there is an effect or relationship when, in reality, there is none.
11
Q
What determines the rate of Type 1 error?
A
The rate of Type 1 error is determined by the value of α, which is usually equal to the significance level of the test. The significance level sets the threshold for rejecting the null hypothesis and is typically set at 0.05 or 0.01.
12
Q
What is a Type 2 error in statistics?
A
A Type 2 error occurs when a statistical test fails to reject a false null hypothesis. In other words, it is a false negative result where the test fails to detect an effect or relationship that actually exists.
13
Q
What determines the rate of Type 2 error?
A
The rate of Type 2 error is given by the value of β. It is related to the power of the test, which is the probability of correctly rejecting a false null hypothesis. The power of the test is equal to 1 - β.
14
Q
How are Type 1 and Type 2 errors related?
A
Type 1 and Type 2 errors are inversely related. By decreasing the rate of Type 1 error (α), the rate of Type 2 error (β) typically increases, and vice versa. There is a trade-off between the two, and researchers must consider the acceptable levels of both errors based on the context and goals of their study.
15
Q
What is incidence in relation to a health condition?
A
Incidence refers to the number of new cases of a health condition that occur within a specific population during a given time period. It represents the rate at which new cases of the condition are diagnosed or reported.
16
Q
How is incidence calculated?
A
Incidence is calculated by dividing the number of new cases of a condition by the population at risk during a specified time period. The result is often expressed as a rate or a percentage.
17
Q
What is prevalence in relation to a health condition?
A
Prevalence refers to the total number of cases of a health condition within a specific population at a particular point in time. It represents the proportion of individuals in the population who have the condition.
18
Q
How is prevalence calculated?
A
Prevalence is calculated by dividing the total number of cases of a condition by the total population during a specific time period. The result is often expressed as a rate or a percentage.
19
Q
What is the relationship between incidence and prevalence?
A
The relationship between incidence and prevalence is influenced by the duration of the condition. In chronic diseases, where the condition persists over a long period, the prevalence is typically much greater than the incidence. In acute diseases, where the condition is short-lived, the prevalence and incidence are often similar. For certain conditions like the common cold, the incidence may be greater than the prevalence due to the high occurrence of new cases within a short time period.
20
Q
What is absolute risk reduction?
A
Absolute risk reduction refers to the decrease in risk associated with a specific activity or treatment compared to a control activity or treatment. It quantifies the difference in risk between the two options. Absolute risk reduction is calculated as the difference in probabilities or rates of a defined endpoint between the two treatments.
21
Q
How is absolute risk reduction calculated?
A
To calculate absolute risk reduction, the probabilities or rates of a defined endpoint for two different treatments are compared. The absolute risk reduction is obtained by subtracting the probability or rate of the endpoint for the control treatment from the probability or rate of the endpoint for the active treatment (pX - pY).
22
Q
What is the Number Needed to Treat (NNT)?
A
The Number Needed to Treat (NNT) is the inverse of absolute risk reduction. It represents the number of patients who need to receive a particular treatment to prevent one event (e.g., morbidity, mortality, or adverse outcome). It provides a measure of the impact of a treatment and helps determine the cost-effectiveness and potential benefits of a treatment option.
22
Q
How is the Number Needed to Treat calculated?
A
The Number Needed to Treat (NNT) is calculated as the reciprocal of the absolute risk reduction. It is obtained by taking the inverse of the difference in probabilities or rates between two treatments. The NNT indicates the number of patients that need to be treated with a specific intervention to prevent one additional event compared to a control group.
22
Why is it important to categorize data correctly before statistical analysis?
Categorizing data correctly before statistical analysis is crucial because it determines the appropriate method of analysis to be used. Different types of data require different statistical tests and techniques. By accurately categorizing the data as nominal, ordinal, interval, or continuous, researchers can select the most suitable statistical approach to draw meaningful conclusions from the data.
22
To calculate absolute risk reduction, the probabilities or rates of a defined endpoint for two different treatments are compared. The absolute risk reduction is obtained by subtracting the probability or rate of the endpoint for the control treatment from the probability or rate of the endpoint for the active treatment (pX - pY).
Absolute risk reduction is important in evaluating the effectiveness of different treatments. It helps quantify the actual reduction in risk associated with a specific treatment compared to an alternative. This information is valuable in making informed decisions about treatment options and assessing the cost versus the potential benefit of a treatment.
23
What is nominal data?
Nominal data refers to data that can be assigned a numerical code, but the code itself is arbitrary and does not convey any inherent order or magnitude. An example of nominal data is categorizing people as alive or dead using codes of 0 or 1.
24
What is ordinal data?
Ordinal data involves numbers that can be used to represent a scale or order. It allows for the ranking or categorization of data based on a specific attribute or characteristic. An example of ordinal data is the measurement of pain severity, where numbers are assigned to indicate the level of pain experienced.
25
What is continuous data?
Continuous data is measured numerically and can take on any real value. It is not limited to specific categories or intervals. Examples of continuous data include measurements such as height, weight, or temperature.
26
What are some common statistical tests used for analyzing different types of data?
For normally distributed data, parametric tests like the T Test are commonly used. Non-normally distributed data requires non-parametric tests such as the Chi Squared test or Mann Whitney U test. The Fisher's exact test is often used for small sample sizes. The paired T Test is appropriate when paired samples are taken from the same individuals, such as before and after an intervention.
27
What is post hoc analysis?
Post hoc analysis refers to the analysis of data after an initial analysis has been performed and unexpected results or differences are observed. It involves examining specific groups or subgroups within the data to identify any patterns or correlations that were not originally anticipated. However, it is important to interpret post hoc analysis with caution as it can increase the likelihood of errors and false rejections of null hypotheses.
28
How can multiple testing be addressed in statistical analysis?
To address the issue of multiple testing, researchers may apply a Bonferroni correction. This correction adjusts the statistical analysis to account for the increased probability of obtaining a statistically significant result by chance when conducting multiple analyses on the same dataset. The Bonferroni correction helps control the overall Type I error rate and maintain the reliability of the statistical findings.
29
What are the three types of consent?
The three types of consent are:
1. Informed consent
2. Expressed consent
3. Implied consent
30
What are the key points of capacity when it comes to consent?
The key points regarding capacity for consent are:
1. Understanding and retaining information
2. Believing the information to be true
3. Weighing the information to make a decision
31
What is qualitative data?
Qualitative data, also known as categorical data, consists of different descriptions or categories of a characteristic. While it may be possible to assign numbers to these categories, they do not have a numerical scale or inherent order. Examples of qualitative data include gender, occupation, or types of food.
32
What are the different consent forms used in the UK NHS?
The different consent forms used in the UK NHS are:
1. Form 1: For competent adults who are able to consent for themselves, even when consciousness may be impaired (e.g., under general anesthesia)
2. Form 2: For an adult consenting on behalf of a child when consciousness is impaired
3. Form 3: For an adult or child when consciousness is not impaired
4. Form 4: For adults who lack capacity to provide informed consent
33
Who can provide consent for minors in the UK?
In the UK, young children and older children who are not deemed Gillick competent cannot provide consent for themselves. The biological mother of the patient can always provide consent. If the parents are married (and the father is the biological father) or if the father is named on the birth certificate (regardless of marital status), the child's father can also provide consent. However, if the parents are not married and the father is not named on the birth certificate, the father cannot provide consent.
34
What is quantitative data?
Quantitative data is data that is associated with numerical values on a numerical scale. It involves measurements or counts that can be expressed as numbers. Quantitative data can be organized to create a distribution curve and allows for various statistical analyses. Examples of quantitative data include height, weight, or test scores.
35
How can quantitative data be analyzed?
Quantitative data can be analyzed by estimating the central tendency using measures such as the mode (most frequently occurring value), median (middle value), and mean (average). The standard deviation provides an estimation of the spread or variability of the data points around the central tendency. Other statistical techniques can also be applied to analyze quantitative data, such as hypothesis testing, correlation analysis, and regression analysis.
36
How is relative risk calculated?
Relative risk is calculated by dividing the experimental event rate (EER) by the control event rate (CER). The formula for relative risk is: RR = EER / CER.
36
What is relative risk (RR)?
Relative risk (RR) is the ratio of the risk in the experimental group (experimental event rate, EER) to the risk in the control group (control event rate, CER). It measures the likelihood or probability of an event occurring in the experimental group compared to the control group.
37
What does a relative risk ratio greater than 1 indicate?
If the relative risk ratio is greater than 1, it indicates that the rate of an event (such as experiencing significant pain relief) is increased in the experimental group compared to the control group.
38
What does a relative risk ratio less than 1 indicate?
If the relative risk ratio is less than 1, it indicates that the rate of an event is decreased in the experimental group compared to the control group.
39
What is relative risk reduction (RRR) or relative risk increase (RRI)?
Relative risk reduction (RRR) or relative risk increase (RRI) is a measure that quantifies the magnitude of the difference in risk between the experimental and control groups. It is calculated by dividing the absolute risk change by the control event rate.
40
How is relative risk reduction (RRR) or relative risk increase (RRI) calculated?
To calculate relative risk reduction (RRR) or relative risk increase (RRI), subtract the control event rate (CER) from the experimental event rate (EER), and then divide the result by the CER. The formula is: RRR/RRI = (EER - CER) / CER.
41
Using the given data, what is the relative risk and relative risk increase?
Based on the given data, the relative risk is 2.4. This means that the rate of experiencing significant pain relief is 2.4 times higher in the paracetamol group compared to the placebo group. The relative risk increase is 140%, indicating a 140% increase in the likelihood of experiencing significant pain relief with paracetamol compared to placebo.
42
What is Type 2 error?
Type 2 error occurs when a test fails to reject a false null hypothesis. It is related to the concept of statistical power. In other words, Type 2 error happens when a test fails to detect a significant effect or relationship that actually exists.
42
What is Type 1 error?
Type 1 error, also known as a false positive, occurs when a test rejects a true null hypothesis. It is equivalent to concluding that there is a significant effect or relationship when there isn't one. The probability of Type 1 error is typically set as the significance level of the test.
43
How is statistical power calculated?
Statistical power is influenced by various factors, such as the sample size, effect size, and the chosen significance level. Power calculations involve estimating these factors to determine the probability of correctly rejecting the null hypothesis. Statistical software or online calculators are commonly used to perform power calculations based on these inputs. Increasing the sample size or effect size, or reducing the significance level, can increase the power of a test.
44
What is statistical power?
Statistical power refers to the probability that a test will correctly reject the null hypothesis when it is false. It represents the ability of a test to detect a real effect or relationship. By increasing the power of a test, the probability of committing a Type 2 error can be reduced. A commonly selected value for statistical power is 0.8, indicating an 80% chance of correctly rejecting a false null hypothesis.
45
What is the relationship between statistical power and Type 2 error?
Statistical power and Type 2 error are inversely related. As the power of a test increases, the probability of committing a Type 2 error decreases. A high-power test is more likely to correctly reject a false null hypothesis, thus minimizing the chance of failing to detect a real effect or relationship.
46
What is the normal distribution?
The normal distribution, also known as the Gaussian distribution or 'bell-shaped' distribution, is a statistical distribution that describes the spread of many biological and clinical measurements. It is symmetrical and has a characteristic bell-shaped curve.
47
What are the properties of the normal distribution?
Symmetry: The mean, mode, and median are all equal.
Percentage within standard deviations: Approximately 68.3% of values lie within 1 standard deviation (SD) of the mean, 95.4% lie within 2 SDs, and 99.7% lie within 3 SDs.
95% confidence interval: Within 1.96 SDs of the mean lies approximately 95% of the sample values. This range is called the 95% confidence interval.
48
What does the standard deviation represent?
The standard deviation (SD) represents the average difference between each observation in a sample and the sample mean. It is a measure of the variability or dispersion of the data points around the mean. The standard deviation is calculated by taking the square root of the variance.
49
What is pre-test probability?
Pre-test probability refers to the proportion of people with a specific disorder in the population at risk before any diagnostic test is conducted. It can be measured as the point prevalence (proportion of people with the disorder at a specific time) or period prevalence (proportion of people with the disorder over a specific time interval). For example, the prevalence of rheumatoid arthritis in the UK is 1%.
50
What is post-test probability?
Post-test probability is the proportion of patients with a particular test result who actually have the target disorder. It is calculated by dividing the post-test odds by the sum of 1 and the post-test odds.
51
How is post-test probability calculated?
To calculate post-test probability, divide the post-test odds by the sum of 1 and the post-test odds. The formula is: Post-test probability = Post-test odds / (1 + Post-test odds).
52
What are pre-test odds?
Pre-test odds represent the odds that a patient has the target disorder before any diagnostic test is performed. It is calculated by dividing the pre-test probability by the complement of the pre-test probability (1 minus the pre-test probability). The formula for pre-test odds is: Pre-test odds = Pre-test probability / (1 - Pre-test probability).
53
What are post-test odds?
Post-test odds represent the odds that a patient has the target disorder after a diagnostic test is conducted. It can be calculated by multiplying the pre-test odds by the likelihood ratio for a positive test result. The formula for post-test odds is: Post-test odds = Pre-test odds x Likelihood ratio (for a positive test result). The likelihood ratio is calculated as the sensitivity divided by 1 minus the specificity of the test.
54
What are the levels of evidence in study design?
Level I: Evidence from meta-analysis of randomized controlled trials.
Level II: Evidence from at least one well-designed controlled trial that is not randomized.
Level III: Evidence from correlation and comparative studies or the use of historical controls.
Level IV: Evidence from case series.
Level V: Expert opinion or evidence founded on basic principles.
Please note that knowledge of the subgroups of levels of evidence is not routinely tested in MRCS Part A
55
What is a randomized controlled trial (RCT)?
A randomized controlled trial is a type of study design where participants are randomly assigned to either an intervention group or a control group. The intervention could be a new treatment or a placebo, while the control group typically receives standard treatment. However, practical or ethical considerations may limit the use of RCTs in certain situations.
55
What is a cohort study?
A cohort study is an observational and prospective study design. It involves selecting two or more groups of individuals based on their exposure to a specific agent, such as a medicine or toxin. These groups are then followed up over time to observe how many develop a certain disease or outcome. The primary outcome measure used in cohort studies is the relative risk. An example of a well-known cohort study is the Framingham Heart Study.
56
How are recommendations graded?
Recommendations are graded based on the strength of the underlying evidence. The grading of recommendation is as follows:
Grade A: Based on evidence from at least one randomized controlled trial (i.e., Level Ia or Ib).
Grade B: Based on evidence from non-randomized controlled trials (i.e., Level IIa, IIb, or III).
Grade C: Based on evidence from a panel of experts (i.e., Level IV).
57
What is a case-control study?
A case-control study is an observational and retrospective study design. In this type of study, patients with a particular condition (cases) are identified and matched with controls. Data is then collected on their past exposure to a potential causal agent for the condition. The primary outcome measure used in case-control studies is the odds ratio. Case-control studies are relatively inexpensive, provide quick results, and are useful for studying rare conditions. However, they are prone to confounding.
58
What is a cross-sectional survey?
A cross-sectional survey provides a "snapshot" of a population at a specific point in time. It is also known as a prevalence study. Cross-sectional surveys provide weak evidence of cause and effect since they only examine the relationship between exposure and outcome at a single point in time. They can be useful for gathering information about the prevalence of a condition or certain factors within a population.
59
What is sensitivity in screening test statistics?
Sensitivity is the proportion of patients with a certain condition who have a positive test result. It is calculated by dividing the true positive (TP) by the sum of true positive (TP) and false negative (FN) results. Sensitivity indicates how well the test can correctly identify individuals with the condition.
60
What is specificity in screening test statistics?
Specificity is the proportion of patients without a certain condition who have a negative test result. It is calculated by dividing the true negative (TN) by the sum of true negative (TN) and false positive (FP) results. Specificity indicates how well the test can correctly identify individuals without the condition.
61
What is positive predictive value?
Positive predictive value is the chance that a patient has the condition if the diagnostic test result is positive. It is calculated by dividing the true positive (TP) by the sum of true positive (TP) and false positive (FP) results. Positive predictive value helps assess the probability of having the condition based on a positive test result.
62
What is negative predictive value?
Negative predictive value is the chance that a patient does not have the condition if the diagnostic test result is negative. It is calculated by dividing the true negative (TN) by the sum of true negative (TN) and false negative (FN) results. Negative predictive value helps assess the probability of not having the condition based on a negative test result.
63
What is the likelihood ratio for a negative test result?
The likelihood ratio for a negative test result is calculated by dividing 1 minus the sensitivity by the specificity. It represents how much the odds of having the disease decrease when a test result is negative.
64
What is the likelihood ratio for a positive test result?
The likelihood ratio for a positive test result is calculated by dividing the sensitivity by 1 minus the specificity. It represents how much the odds of having the disease increase when a test result is positive.
65
Are positive and negative predictive values prevalence dependent?
Yes, positive and negative predictive values are prevalence dependent. This means that their values can vary depending on the prevalence of the condition in the population being tested.
66
Are likelihood ratios prevalence dependent?
No, likelihood ratios are not prevalence dependent. This means that their values are not influenced by the prevalence of the condition in the population being tested.
67
What is absolute risk reduction?
Absolute risk reduction is the measure of the decrease in risk associated with a specific activity or treatment compared to a control activity or treatment. It is calculated as the difference between the probabilities of a defined endpoint, such as 5-year survival, for two different treatments. For example, if the probabilities of the endpoint for surgical resection (X) and watchful waiting (Y) for prostate cancer are known as pX and pY, respectively, the absolute risk reduction is calculated as pX - pY. It provides valuable information about the effectiveness of treatments.
68
What is the Number Needed to Treat (NNT)?
The Number Needed to Treat (NNT) is the inverse of the absolute risk reduction. It represents the number of patients that would need to receive a specific treatment in order to prevent one event, such as a positive outcome or occurrence of a disease. The NNT is calculated as the absolute difference between the two treatments. The NNT helps determine the balance between the cost and benefit of various treatments by providing an estimate of the number of patients who need to be treated to achieve a desired outcome.
69
What are cluster randomised controlled trials?
Cluster randomised controlled trials are a type of study design where groups, or clusters, of participants are randomly assigned to different interventions or control conditions, rather than randomising individuals. This design is often used when individual randomisation is not feasible or practical, such as in community or organizational settings. By randomising clusters, the trials aim to avoid cross-contamination among participants within the same cluster. Clusters can be defined as schools, hospitals, neighborhoods, or any other group of individuals.
70
Why are participants in the same cluster more likely to respond similarly in cluster randomised controlled trials?
Participants within the same cluster in cluster randomised controlled trials are more likely to respond in a similar fashion due to shared characteristics, environment, or contextual factors. For example, if a cluster represents a school, students within the same school may have similar educational experiences, resources, or influences that can affect their response to the intervention or control condition. This similarity in response within clusters needs to be considered when analyzing the trial results.
71
What is the risk of unit of analysis error in cluster randomised controlled trials?
Cluster randomised controlled trials are at a higher risk of unit of analysis error. This means that the studies should be analyzed and interpreted at the cluster level rather than at the individual level. Failing to account for the clustering effect can lead to a higher false positive rate, which means that the observed effects may be erroneously attributed to the intervention when they are actually due to the clustering of participants within the same cluster.
72
Is it possible to adjust for clustering in statistical analyses of cluster randomised controlled trials?
Yes, it is possible to adjust for clustering in statistical analyses of cluster randomised controlled trials. Various statistical methods, such as multilevel modeling or generalized estimating equations, can be used to account for the clustering effect and obtain accurate estimates of treatment effects. These methods allow for the appropriate adjustment of standard errors and the consideration of within-cluster correlation when analyzing the trial data. Adjusting for clustering helps mitigate the risk of unit of analysis error and allows for more accurate interpretation of the trial results.
73
What is sensitivity in the context of screening tests?
Sensitivity is the proportion of true positives that are correctly identified by a screening test. It indicates how well the test can detect individuals who have the disease or condition being screened for. Sensitivity is calculated by dividing the number of true positives by the sum of true positives and false negatives.
74
What is specificity in the context of screening tests?
Specificity is the proportion of true negatives that are correctly identified by a screening test. It indicates how well the test can correctly identify individuals who do not have the disease or condition being screened for. Specificity is calculated by dividing the number of true negatives by the sum of true negatives and false positives.
75
What is positive predictive value?
Positive predictive value is the proportion of individuals who have a positive test result and actually have the disease or condition being screened for. It provides an estimate of the likelihood that a positive test result is indicative of the presence of the disease. Positive predictive value is dependent on both sensitivity and specificity, as well as the prevalence of the disease in the population being tested.
76
What is negative predictive value?
Negative predictive value is the proportion of individuals who test negative and do not have the disease or condition being screened for. It provides an estimate of the likelihood that a negative test result accurately indicates the absence of the disease. Negative predictive value is also dependent on sensitivity, specificity, and disease prevalence.
77
Are predictive values dependent on the prevalence of the disease?
Yes, predictive values, both positive and negative, are dependent on the prevalence of the disease in the population being screened. The prevalence of the disease affects the likelihood of true positives and true negatives, which in turn influences the positive predictive value and negative predictive value. Therefore, it is important to consider disease prevalence when interpreting these values.
78
What is the likelihood ratio for a positive test result?
The likelihood ratio for a positive test result is calculated by dividing the sensitivity of the test by 1 minus the specificity. It represents how much more likely individuals with the disease are to have a positive test result compared to those without the disease. Likelihood ratios provide information about the diagnostic accuracy of a test and are not influenced by disease prevalence.
79
What is the likelihood ratio for a negative test result?
The likelihood ratio for a negative test result is calculated by dividing 1 minus the sensitivity by the specificity. It represents how much more likely individuals without the disease are to have a negative test result compared to those with the disease. Likelihood ratios provide information about the diagnostic accuracy of a test and are not influenced by disease prevalence.
80
What is a null hypothesis in significance tests?
A null hypothesis (H0) in significance tests states that two treatments or conditions are equally effective, and it is negatively phrased. It serves as a baseline assumption that there is no significant difference between the treatments being compared. For example, in a study comparing the prevalence of colorectal cancer in patients taking low-dose aspirin versus those who are not, the null hypothesis would state that there is no difference in the prevalence between the two groups.
81
What is the p-value in significance tests?
The p-value is a measure of the probability of obtaining a result as extreme as the one observed, assuming that the null hypothesis is true. It indicates how likely the observed data are if the null hypothesis is correct. A lower p-value suggests stronger evidence against the null hypothesis. The p-value is also equal to the chance of making a type I error, which is rejecting the null hypothesis when it is actually true.
81
What is an alternative hypothesis in significance tests?
The alternative hypothesis (H1) is the opposite of the null hypothesis. It suggests that there is a difference between the two treatments or conditions being compared. In the example of the prevalence of colorectal cancer, the alternative hypothesis would state that there is a difference in the prevalence between patients taking low-dose aspirin and those who are not.
82
What are the two types of errors that can occur in significance tests?
The two types of errors that can occur in significance tests are type I and type II errors. A type I error occurs when the null hypothesis is rejected when it is actually true, resulting in a false positive. This error is determined against a preset significance level (alpha). On the other hand, a type II error occurs when the null hypothesis is accepted when it is actually false, failing to detect a difference that exists, resulting in a false negative. The probability of making a type II error is termed beta.
83
What is the power of a study in significance tests?
The power of a study is the probability of correctly rejecting the null hypothesis when it is false. In other words, it is the ability of a study to detect a difference when it truly exists. Power is equal to 1 minus the probability of a type II error (beta). Increasing the sample size of a study can increase its power, as it improves the ability to detect true differences between treatments or conditions.
84
What is a Forest plot?
A graphical display that shows the relative strength of treatment effects in multiple scientific studies addressing the same question.
85
How can the power of a study be increased?
The power of a study can be increased by increasing the sample size. A larger sample size improves the study's ability to detect true differences between treatments or conditions. By having more participants, the study becomes more representative of the population, reducing the likelihood of missing important effects. Increasing power also helps minimize the risk of type II errors, which occur when a difference exists but is not detected.
86
What does the size of each square in a Forest plot represent?
The weight of the study in the meta-analysis.
87
When are Forest plots often used?
To visually present meta-analyses of randomized controlled trials.
88
Why are confidence intervals symmetrical in Forest plots?
To ensure that undue emphasis is not given to odds ratios greater than 1 compared to those less than 1.
89
How is the overall measure of effect represented in a Forest plot?
As a vertical line.
89
What does a vertical line representing no effect indicate in a Forest plot?
That the effect size is not significantly different from no effect.
90
What does the diamond shape in a Forest plot represent?
The meta-analyzed measure of effect, with the lateral points indicating confidence intervals for this estimate.
91
What does it mean when the confidence intervals overlap with the line of no effect?
It shows that the effect sizes of individual studies do not differ significantly from no effect at the given level of confidence.
92
What does it mean when the points of the diamond overlap the line of no effect?
The overall meta-analyzed result is not significantly different from no effect at the given level of confidence.
93
What is another name for the normal distribution?
Gaussian distribution or 'bell-shaped' distribution.
94
How is the 95% confidence interval defined in the normal distribution?
It is the range from the mean minus 1.96 times the SD to the mean plus 1.96 times the SD.
94
What does the normal distribution describe?
The spread of many biological and clinical measurements.
95
What are the properties of the normal distribution?
Symmetry (mean = mode = median), 68.3% of values within 1 standard deviation (SD) of the mean, 95.4% within 2 SD, and 99.7% within 3 SD.
96
What does the standard deviation (SD) represent?
The average difference between each observation in a sample and the sample mean.
97
How can the standard deviation be calculated?
By taking the square root of the variance.
