Reference: Mohammad, A., et al., Prevalence of fibromyalgia among patients with chronic hepatitis C infection: relationship to viral characteristics and quality of life. J Clin Gastroenterol, 2012. 46(5): p. 407-12.
A study out of Ireland on the risk factors for fibromylagia syndrome (FBS) among chronic Hepatitis C patients.
Brief Summary: Many patients with chronic Hepatitis C can manifest many different types of symptoms such as fatigue, myalgia, and fibromylagia syndrome (FBS). FBS is a medical disorder characterized with widespread muscle, tissue or joint pain. The main objective of this study was to identify the observational factors which are associated with patients who have FBS vs. no-FBS in chronic Hepatitis C patients. A total of 185 Hepatitis C patients were recruuited and a wide variety of observatioal factors were collected and recorded such as gender, age, pain intensity, functional impairment, etc.
Results: The authors found the following factors to be risk factors for FBS among the patients with chronic Hepatitis C patients: age 45 years or more, female sex, living alone, history of depression, acquisition of HCV through blood transfusion, and presence of HCV genotype 1.
Implications for Practice: Chronic Hepatitis C Patients who show the corresponding risk factors may have a higher liklihood of having FBS.
Discussion: As the authors mentioned, the pathogenesis of FBS is not completely understood. Thus, observational studies like this which identify risk factors can be used to further elucidate the cause of FBS. I obviously have not had the opportunity to read up on past papers related to this subject, but just from the author’s discussion section, it would seem as if there is a strong interaction between several variables – both environmental and genetic which can cause FBS. There are just so many follow-up studies that could be done on a project such as this.
Commentary on Statistics and Study Design: I have some constructive suggestions for the authros related to the statistics and study design of the investigation. The main objective of the investigation was to identify risk factors for patients who either have FBS or do not have FBS. Since the main desire was to identify independent risk factors, conducting a univariate logistic regression model fulfills this need. However, it would also be helpful to identify risk factors in a dependent fashion. In order to do this, the author’s could have constructed a multi-variate logisticregression model with the outcome variable (as before) being whether the patient had FBS or not, and then the predictor variables would have been all the risk factors. By doing this, the investigator could determine whether a given risk factor (or variable) is statistically associated with the outcome variable while controlling for (or ‘keeping constant’) all other variables in the investigation. This could help the investigator in narrowing down the risk factors which are most associataed with the outcome variable, and this is information which could be very useful. For instance, gender was found to be independantly associated with the outcome variable (FBS vs. no-FBS) in the univariate analysis. However, gender may not be associated with the outcome variable while controlling for – say – blood transfusion. In other words, blood transfusion has a much stronger association with the outcome variable than gender, and this would be useful to know. In this instance, gender is not really associated with the outcome variable – it may only been because more males than females happened to have FBS in the recruitment than males. So, the author’s could have done a multi-variate analysis and only reported those risk factors which were associated with the outcome variable in a dependant fashion. In short, whenever there are multiple predictor variables in a problem, you always want to run a multi-variate regression problem (for the reasons just stated) instead of a univariate approach. Also, by using a multi-variate regression problem, one could have tested the effect of various interactions among the different variables (ask me if you don’t know what this is).
Also, the author’s identified 4 different genotypes among the HC patients. However, in the final analysis, the investigators only made a comparative difference between genotype 1 vs. every other genotypes. It may have been helpful to do a comparative analysis between all the genotypes simultaneously (ex. genotypes 2 vs. 3, 2 vs. 4, 1 vs. 4), and there are statistical techniques to do this. It seems as if there were plenty of data samples for genotypes 2 through 4 to do this. Using the multi-variate regression idea, the investigator could have coded the genotypes variable as a qualitiatve class variable, and this would have allowed the comparison of all 4 groups against themselves (again, ask if you don’t know how to do this). This could have been really interesting to look at.
A big thanks for our friends from Ireland for running this study.
Source- http://thegastroenterologyblog.blogspot.com/2012/05/risk-factors-for-fibromylagia-syndrome.html
No comments:
Post a Comment