Statistics! Week 8, basic research skills

 

 

Struggling to say the least. I will attempt to look at a quantitative paper from a data collection, statistical analysis and interpretation point of view. I have to read the ‘methods’, analysis’ and ‘results’ sections then I will ask the following questions: –

What outcome measures did they use i.e. how did they measure and collect their data? The data was collected from questionnaires at varying stages in the study up to three years, they did not give up. No one dropped out, so they say, this seems surprising

Do they talk about reliability, validity and sensitivity? No. Small sample size. They did not do a sample size calculation.

Are the measurement scales they used reliable, do they give references for them? Yes, they are all well known and referenced scales.

Look at the summary baseline data first. This should be in a table form. Are there any systematic differences between the groups? The mean differences (standard deviations) between the groups are displayed transparently they have been randomly selected from 24 volunteer women from Oslo form 5 different work places – I see a reliability problem here.

Do they discuss any assumptions they have made about the data? They assume it has a normal distribution. They use a one tailed test (uni-variate). I cannot find ‘Student’s matched pairs’ test in commonly used statistics. I can find the paired t-test and a student’s t-test which seems to fit with what they are trying to achieve – measure the mean differences between matched pairs for any significant difference in the mean scores of nominal, normal distribution data. Did they use a parametric test for skewed data and hence the positive outcome? Did they assume a positive outcome and that’s why they used a one tailed test? No, uni-variate means they only look at one variable and do not try to explain it in the light of another. The use another test to do this to see any correlation between variables – these are the researches assumptions of important variables- they used a ‘Pearson’s correlation coefficient ‘ for this which appears appropriate, but who knows!

Do they say which tests they have used and why? Yes, the tests are mentioned as above but no justification was given. They do give narrow confidence intervals regarding the treatment effect rather than just saying it was a significant difference. Do they tell me what the mean difference was – I’m not sure?

Do the tests seem reasonable? I think so?

Are there outcomes missing? No

Do the conclusions follow from the data presented? Yes, I think their conclusions are honest and transparent noting small sample size. They do not mention that this is more than likely not a representative sample people with neck and shoulder pain generally. They make tentative conclusions based on the data.

To summarise, small sample size which was randomised in to 2 matched groups, well matched, and table clear, including standard deviations on nine characteristics. Control group received scam acupuncture, the same clinician did all the treatments and tried to be the same with everyone, bearing in mind it was impossible to blind the clinician so it was a single blind experiment – they did not see any methodological problem for this claiming to have controlled it well enough.They presumed a normal distribution data and did a parametric test, this may be why they achieved a significant difference in groups. I think it was a 2 tailed test but not explicit. The results of the different rating scales were clearly shown in graph form between the groups. Tentative conclusions only were made based on the data -acupuncture ‘may’ improve neck and shoulder pain, they did not say that they were unable to generalise to wider population which I don’t think they can do with a sample size of 24. They did what they said they were going to do although I’m not sure why they did the correlations.

Paper discussed is below, complete with link if anyone else would like to tackle it!

Effect of intensive acupuncture on pain-related social and psychological variables for women with chronic neck and shoulder pain (2005)

Dong He, Arne T Høstmark, Kaj Bo Veiersted, Jon Ingulf Medbø

Acupuncture in Medicine 23 (2) 52-61

https://www.researchgate.net/profile/Kaj_Veiersted/publication/8548064_Effect_of_acupuncture_treatment_on_chronic_neck_and_shoulder_pain_in_sedentary_female_workers_a_6-month_and_3-year_follow-up_study._Pain/links/0fcfd511d295caf008000000.pdf

Having sorted that, the next job is an assignment that needs to be in 2.12.16, 1,500 words – how I am going to carry out my scoping review.

I understand what is needed of me, explain what I am going to search for, in which data bases, my inclusion and exclusion criteria and how am I going to analyse it. My problem is the search terms, what do I want to know. I will think out loud, it may help – reflection in action, Rolfe, G. (1998). The theory-practice gap in nursing: from research-based practice to practitioner-based research. Journal of Advanced Nursing, 28(3), 672–679. https://doi.org/10.1046/j.1365-2648.1998.00806.x

My ontological and epistemological researcher identity is from the constructionist/ interpretivist  end of the continuum. We build out own knowledge and reality from our interactions with society, we make our own reality form our individual experience of the world, it it different for each of us. we construct our own reality and learn from our experience.

Having said that, as a researcher I would like to know more about the results I am getting from the combined therapies of hypnotherapy and western medical acupuncture. There are so many variables to consider in a whole consultation and most of them are implicit. It may be more useful to study the patients perspective of the experience to gain more information about what is going on.

My problem is there is not much or even no good research out there, that I can find, about the subject (I have tried many and varied search terms). Although I have highlighted a potential gap this may be because there is very little to gain from research in to such a small area, not many people practice this therapy so how useful is it to society to study it? It may be useful clinically and personally but not to a wider population which is the whole point of research in the first place!

I will utilise the PICO model to highlight my search possiblities and my confusion as to which area I should look at:

P – Population – clients, who’s clients, no need to worry about this at the moment, human, adults most likely

I – intervention – Hypnoacupuncture, hynopuncture, hypnosis-acupuncture, hypnotherapy and acupuncture, or do I just study one or the other in depth, from the patient’s perspective.

C – comparison – compared to what they were like before they attended

O – outcome, how do they feel now?

I will put this out to the forum to see what response I receive.

Reply below: –

Looking briefly at the searches your completed for your FA2 assignment, it seems to me that there is sufficient material to look at any of the questions you are considering for your Scoping Review – so conditions-wise, anxiety seems to come through, but there also seem to be papers in the area of combining acu with hypnoacu more generally. As that is an area you are particularly keen to look into in more depth, maybe pursue this at a general level, or with with a focus on the patient experience as per your FA2 RQ.

You do not need ‘lots’ of papers for your Scoping Review. What you do need to do is go through a structured process to define your RQ (as you’ve have started to demonstrate in your post), identify a search strategy, complete the search, select your papers against your PICO (or other) criteria, critically appraise the papers and extract the relevant data, discuss your findings and present your thoughts as conclusions/recommendations. So you just need enough papers to be able to do this.

Next week ethics, assignment due in tomorrow – no pressure!

 

Leave a Reply

Your email address will not be published.