Chapter 27 – Evidence Management: Data Analysis and Clinical Trials Structured SBA

Abstract

Evidence Management: Data Analysis and Clinical Trials Structured SBA Questions

Chapter 27 Evidence Management: Data Analysis and Clinical Trials Structured SBA

Munier Hossain

Evidence Management: Data Analysis and Clinical Trials Structured SBA Questions

1. With regard to normal distribution, which of the following is true?

50% of samples in the dataset lie above the mean value

Mean ±1 standard deviation would capture 95% of the data

Mean and interquartile range are the best parameters to describe the data

Normal distribution is useful to plot categorical variables

There are minor differences between mean, median and mode

2. A paired sample t test is appropriate for which of the following situations?

Comparing bone mineral density between those with and those without ankylosing spondylitis

Investigating pain relief on a visual analogue scale before and after hip arthroplasty

Investigating proportion of pain relief on a visual analogue scale before and after hip arthroplasty

Comparing patient satisfaction with radiological outcome after ankle fracture fixation

Comparing visual analogue scale scores after total hip arthroplasty and hip hemiarthroplasty following hip fracture

3. Researchers investigated pain relief and functional outcome before and after resection arthroplasty of the shoulder. The level of pain was significantly decreased (t test, p < 0.001).

Which of the following is true?

Pain relief following resection arthroplasty was noticeable

Resection arthroplasty can be recommended as a treatment option

Resection arthroplasty has a strong effect on pain relief

The very low p value is indicative of very low bias of the study

There is <1% probability that the observed differences in pain relief is due to chance

4. Which of the following is true?

A wide confidence interval gives more precise estimate

A small trial would result in a narrow confidence interval

The range of the confidence interval can be used to test the null hypothesis

If the confidence intervals of two different interventions do not overlap, this would suggest that there is a clinically significant difference in treatment effect

A 95% confidence interval is always used in research

5. Which of the following is true?

P = 0.001 means the intervention has a stronger effect than p = 0.01

P < 0.05 suggests that the treatment effect was clinically significant

The lower the p value, the greater the strength of evidence against the null hypothesis

Threshold for P value is always <0.05

P value is constant and is not affected by multiple statistical testing

6. Researchers were interested to investigate risk factors for deep infection in patients with a hip arthroplasty. They employed a case-control design in which 47 consecutive patients with a deeply infected hip arthroplasty were compared with 200 randomly selected patients and operated on in the same hospital and period of time, with no deep infection in their hip arthroplasty along follow-up.

Which of the following is true?

The study was free of allocation bias

The study was free of assessor bias

The study was free of confounding

The study was free of selection bias

This was a retrospective study

7. An orthopaedic surgeon is interested to investigate the efficacy of a new treatment for knee pain. He wishes to conduct a randomised double-blind trial.

Which of the following is true?

Appropriate randomisation can only be performed using computer-generated random numbers

Double-blind design would help to minimise observer bias

Effect of confounding factors can be minimised if randomisation is adequately performed

Randomisation would be adequate if patients are allocated according to the day of the week of their presentation to the hospital

The study would be free from selection bias

8. A placebo-controlled trial was conducted to compare botulinum toxin to placebo in management of tennis elbow. Quality of life, pain and grip strength were assessed pre-injection and at 3 months’ post-injection in both affected and non-affected arms. Following BTX-A treatment, patients had average 19% improvement in grip strength in the affected arm compared to an average 2% for placebo, (p = 0.08, 95% CI −2.31, 35.64). Both groups showed similar improvements in pain assessment and also in quality of life.

Which of the following is true?

Botulinum toxin group did not benefit from the placebo effect

Improvement in the placebo arm was not due to natural variation of the disease

Placebo effects also modulated the treatment response

Treatment response was solely due to the intervention

Treatment response is the difference in grip strength between the intervention and placebo

9. Researchers investigated the efficacy of bandage vs ankle support for management of ankle instability. Mean MOXFQ score was noted before and after the procedure. Data were analysed using a t test.

Which of the following is true?

Independent sample t test would be an appropriate test for this trial

The authors were required to be satisfied that homogeneity of variance existed before undertaking the t test

Data from the trial can be appropriately displayed in a box plot

Mann-Whitney U test is an appropriate test for data with binary distribution

The MOXFQ score was normally distributed in this population

10. Data distribution of a variable was displayed using the chart below (Figure 27.1).

Which of the following statements is true?

Figure 27.1 Data distribution

A box plot is a suitable chart to describe the data

If we calculate 95% confidence interval of the sample estimate, this would suitably exclude 5% of the dataset

Mean value is to the right of median value in Figure 27.1

One could utilise a student’s t test to test for null hypothesis

The dataset here can be adequately described by the sample mean and confidence interval

11. Researchers were interested in assessing the effectiveness of physiotherapy compared to watchful waiting in patients with plantar fasciitis. The primary outcome was self-reported heel pain using a 0–10mm visual analogue scale (VAS). The minimal clinically important difference (MCID) was 2mm. The authors concluded that physiotherapy was significantly highly likely to improve pain compared to watchful waiting (p = 0.02, mean VAS for physiotherapy group was 4.2mm, mean VAS for watchful waiting group was 5.3mm).

Which of the following statements is true?

Majority of patients who had physiotherapy reported a VAS score of 4.2mm

The difference in pain relief between physiotherapy and watchful waiting was clinically significant

There is a 2% probability that physiotherapy is more effective compared to watchful waiting for treatment of frozen shoulder

There is a 2% probability that the difference in pain relief between physiotherapy compared to watchful waiting was due to chance

A larger sample would have meant that the observed differences might reach clinical significance

12. The abstract of a paper reported ‘the mean difference in pain relief between the intervention group and the control group as calculated using a 0–100mm VAS was 13.2mm (95% CI –8.1mm to 31.1mm)’.

Which of the following statements is true?

95% of the control group reported –8.1mm to 31.1mm pain relief

95% of the intervention group reported –8.1 to 31.1mm pain relief

99% CI would give us a more precise estimate

A p value would have been useful to calculate statistical significance

The observed difference in pain relief between the intervention group and the control group was probably due to chance

13. Researchers compared the vertebral bone strength between control group and those treated with balloon bone tamp or polymer bone stamp. Results are presented in Figure 27.2.

Figure 27.2 Box plot comparing vertebral body strength between three groups. Figure courtesy of Seel EH, Davies EM. A biomechanical comparison of kyphoplasty using a balloon bone tamp versus an expandable polymer bone tamp in a deer spine model. J Bone Joint Surg Br. 2007 Feb;89(2):253-257.

Which of the following is true?

Data presented in this graph had a parametric distribution

Pre-augmentation control group had most outliers

The graph shows mean, interquartile range and extreme values

The graph shows median, interquartile range and extreme values

There was statistically significant difference in bone strength between the balloon tamp group and the polymer bone tamp group

14. Researchers compared heterotopic ossification (HO) and range of movement between elective and trauma elbow arthroplasty. They found that the presence of HO did not significantly affect elbow range of movement within the trauma or elective groups (elective arthroplasty, Mann–Whitney U test, p = 0.070; trauma arthroplasty, p = 0.370, Mann–Whitney U test).

Which of the following is true?

If the data were subjected to logarithmic transformation, this would have allowed the researchers to analyse data via a paired sample t test

Median value is the appropriate measure of central tendency for the above data

There is at least 7% probability that the difference in range of motion observed between patients with or without HO after trauma elbow arthroplasty is due to chance or random variation

There is no difference in range of movement between elective and trauma elbow arthroplasty

While analysing data, researchers made assumptions regarding the distribution of the data

15. Researchers investigated if poor bone quality predisposes to loosening after total hip arthroplasty (THA). Patients with loose hip replacements had more pain and were more likely to have a history of fragility fracture, narrower femoral cortices and lower periprosthetic or lumbar spine bone mineral density (all t test, p < 0.01). They also had vitamin D deficiency (t test, p = 0.31).

Which of the following is true?

It is likely that the difference in vitamin D level between those with and without loosening after THA was due to random variation

Paired sample t test was performed

There was a 1% probability that someone with narrow femoral cortices will develop loosening after THA

There was a 31% difference in vitamin D level between patients with or without a loose THA

There was no need to assess the distribution of the variables

16. Researchers investigated if poor bone quality predisposes to loosening after total hip arthroplasty (THA). Clinical, radiological, biomechanical and bone mineral density indicators of bone quality were assessed. Patients with loose replacements also tended to be smokers (Chi square test, p = 0.08).

Which of the following is true?

Researchers compared a continuous variable between the two groups

Researchers compared a binary variable between the two groups

The Fisher exact test is not a suitable test for this cohort

There is around 8% difference in smoking status between those with and without stem loosening

The results suggest that stem loosening and smoking status are not independent variables

17. Researchers were interested in assessing the effectiveness of a new implant in treating patients with shoulder fracture. The control group was treated with the PHILOS plate. The trial was randomised, and the assessors were appropriately blinded. There were 350 participants in each group. There was loss to follow-up; some patients also did not receive the intended implants. Therefore, only 269 patients in the control group and 250 patients in the intervention group completed the trial protocol. The results were analysed according to intention to treat analysis.

Which of the following statements is true?

The trial was at risk of observer bias

Intention to treat analysis is the best analysis option to minimise the effect of confounding

The trial was at risk of allocation bias

Since randomisation was appropriately performed, bias was eliminated from the trial

Per protocol analysis is the best analysis option due to crossover

18. Researchers reviewed the outcome of the treatment by impaction grafting of periprosthetic femoral fractures. They found that odds ratio of union when treated by impaction grafting and a long stem compared to impaction grafting and a short stem was 5.5 (95% CI 1.54 to 19.6; p = 0.009).

Which of the following statements is true?

Patients treated with impaction grafting and long stem may be up to 20 times more likely to have union than those with impaction grafting and a short stem

Odds of union for a group was the rate of union among the population at risk

Odds ratio of union was calculated by dividing the rate of union in the long stem group vs the rate of union in the short stem group

Impaction grafting was not beneficial for bony union

Long stem implant was the critical intervention that helped in bony union

19. Researchers compared pain and functional outcome after total knee replacement between pessimist and non-pessimist patients. At 2 years after surgery, the odds ratio of pain among pessimists was 2.21 (95% CI 1.12 to 4.35; p = 0.02), odds ratio at 5 years was 1.21 (95% CI 0.51 to 2.83; p = 0.67).

Which of the following is true?

Pessimists reported less pain relief over the study period compared to non-pessimists

Short-term difference in pain relief was not sustained over the long term

Pessimist group reported twice the severity of pain compared to the non-pessimist group at 2 years

Odds ratio of pain was the ratio of patients in the pessimist group reporting pain as compared to the patients in the non-pessimist group reporting pain

Odds of pain in the pessimist group was a reflection of the number of pessimist participants reporting pain among all pessimist participants

20. In a study assessing the 1-year risk of symptomatic venous thromboembolism (VTE) following primary total hip replacement (THR), researchers found that among the patients undergoing THR, the risk of symptomatic VTE was 0.79% between 0 and 90 days after surgery and 0.29% between 91 and 365 days after surgery. In the comparison cohort, the corresponding risks were 0.05% and 0.12%, respectively. The adjusted relative risks of symptomatic VTE among patients undergoing THR were 15.84 (95% CI 13.12 to 19.12) during the first 90 days after surgery and 2.41 (95% CI 2.04 to 2.85) during 91 to 365 days after surgery, compared with the comparison cohort.

Which of the following is true?

Absolute risk of symptomatic VTE was large in the first 90 days after surgery

Symptomatic VTE was between 13 and 19 times more severe in the THR group initially

Patients who had THR may be 241% more likely to develop symptomatic VTE compared to the control group after the first 90 days

The probability of symptomatic VTE for a patient in the THR group was 15.84% in the first 90 days after surgery

When discussing options of chemical thromboprophylaxis with a patient undergoing THR, the most appropriate figure to quote to a patient would be the increased relative risk of 15.84 at the first 90 days and 2.41 up to a year after surgery

21. Researchers in a study found that Gram stain had a sensitivity of 65% and specificity of 40% for diagnosing septic arthritis of the knee.

Which of the following statements is true?

65% patients identified as having a septic arthritis had actual septic arthritis

Only gold members can continue reading. Log In or Register to continue