1) Proposed Research Title:

Reducing Bias in Observational (non-interventional) Studies for Obesity and Diabetes Management.

2) Introduction:

While in experimental trials a direct intervention can be made, observational studies only allow passive observation of the occurring events, making them susceptible to a high degree of bias. The high risk of bias associated with observational studies makes it difficult to establish a causal inference and may lead to inaccurate interpretation of results.

There is a need to reduce associated bias in observational studies being undertaken.

The reduction of bias should first identify the possible sources and types of bias that may be associated with a particular context [1].

The three main types of biases in observational studies include selection bias, confounding bias and information bias. A selection bias leads to the lack of inclusion of some participants into the study, therefore reducing the representativeness of the research. Confounding bias results when the statistical analysis does not consider all the relevant effects. Information bias arises from factors such as misclassification and false reporting of the findings [2].

 

3) Research Questions and Objectives: 

This research aims to identify ways that can be used to reduce the level of bias in observational studies, on the effectiveness of the lifestyle changes, in particular exercise, for the management of obesity in patients with type 2 diabetes.

  • How is the quality of research achieved in observational studies?
  • What are some sources of biases in observational studies relating to the management of diabetes and obesity?
  • How can the sources of biases be reduced in observational studies for managing diabetes and obesity?

The objectives that are pursued in the research questions include:

  • To determine ways for ensuring quality in observational studies.
  • To explore the sources of biases in observational studies relating to the management of diabetes and obesity.
  • To recommend measures that can be used in the reduction of the identified biases of the selected studies.

 

Justification and Rationale:

Regardless of the existence of numerous observational studies on the management and prevention of obesity and diabetes, the data presented may be biased due to the subjective nature of observational studies [3]. For instance, selection bias can occur when the researcher fails to include some participants in the research, therefore reducing the representativeness of the process.

Since obesity and diabetes are significantly affected by environmental and social factors, the biases from an observational study may lead to unreliable conclusions regarding the management measures adopted. This study is therefore justified in the identification of the errors in such studies. The findings of this study will contribute towards the improvement of the methodological processes in obesity and diabetes studies.

 

4) Short Literature Review:

According to the WHO the prevalence of obesity and diabetes has increased dramatically in the past three decades in all countries regardless of their income levels. The number of individuals with diabetes increased from 108 million in 1980 to 422 million in 2014. Similarly, the prevalence of diabetes in adults rose from 4.7% in 1980 to 8.5% in 2014. The prevalence of obesity has also increased in the past 50 years to pandemic levels globally [4]. Obesity presents a major health challenge leading to a substantial increase in the risk of diseases like type 2 diabetes, obstructive sleep apnoea, myocardial infarction and stroke.

Accordingly, the management of the two conditions is of the utmost importance in the global health sector. The reduction of obesity and diabetes burden is based on the combination of individual interventions and changes in the environment and society [5].

Therefore, an understanding of the influence of the personal behaviours on the management of the two conditions is critical in establishing the relative effectiveness of the appropriate strategies. Some of the lifestyle strategies that are used in the prevention and treatment for both diabetes and obesity include the maintenance of a healthy diet, regular physical activities, and avoiding tobacco and alcohol use.

5) Data Collection& Analysis by Systematic Review:

The proposed study involves a comparison of observational studies against randomised control trial (RCTs) regarding the measures of managing obesity via exercise and type 2 diabetes. Following a systematic review all relevant observational studies will be compared against  RCTs.

The inclusion and exclusion criteria will be based on factors such as year of publication (to be published within the last 5 years) and relevance to the topic of discussion.  Therefore the studies excluded will be those published more than five years ago and do not address the management of obesity and diabetes through changes in lifestyle.

Additionally, studies that focus on the changes in lifestyle habits towards the management of other diseases other than obesity and diabetes are excluded.

The study population is represented by adults over 35 years. The data collection is quantitative and qualitative, including questionnaires and statistical diagrams to represent the results of the study.

The main databases that will used for the search are Clintrials.gov, MEDLINE and PubMed with the search restricted to studies published in the last five years. The search strategy retrieves the studies that are then assessed for eligibility and inclusion.

The search terms used in the study include ‘observational studies’, ‘non-interventional studies’, ‘random clinical trials’, ‘interventional studies’, ‘diabetes’, ‘obesity’, ‘management’, ‘exercise’ and ‘lifestyle changes’.

6) Project Management Timetable:

Month Activity Completion
Mar 2020 Database Search, Systematic Review and Final Selection of Studies for Analyses
Mar 2020 Analyse Studies
Mar 2020 Draft Dissertation
Apr 2020 Finalise Dissertation
May 2020 Submit Final Draft for comment
Jun 2020 Revise and Submit Dissertation Project

 

7) Management of Risks to Project:

The main risk to this systematic review project is the risk of bias assessment of the validity of the included studies. The risk of bias entails the overestimation and underestimation of the effect of the intervention in the included studies.

The risk of bias will be ensured by enhancing the validity of the included studies through the adoption of the inclusion and exclusion criteria to enhance the quality of the evidence.

 

8) Ethical considerations:

No ethics submission is needed

The main ethical factors in the project relate to the considerations undertaken in the selected studies as the randomised control trials involve human subjects. Consequently, the included studies are those with ethical approvals.

9) Financial costs: NONE

10) Key References:

  1. Roglic G. WHO Global report on diabetes: A summary. International Journal of Non communicable Diseases. 2016 Apr 1; 1(1):3.
  2. Sharma M, Nazareth I, Petersen I. Observational studies of treatment effectiveness: worthwhile or worthless?. Clinical epidemiology. 2019; 11:35.
  3. Mokdad AH, Ford ES, Bowman BA, Dietz WH, Vinicor F, Bales VS, Marks JS. Prevalence of obesity, diabetes, and obesity-related health risk factors, 2001. Jama. 2003 Jan 1; 289(1):76-9.
  4. Dietz WH, Baur LA, Hall K, Puhl RM, Taveras EM, Uauy R, Kopelman P. Management of obesity: improvement of health-care training and systems for prevention and care. The Lancet. 2015 Jun 20; 385(9986):2521-33.
  5. Naimi TS, Stockwell T, Zhao J, Xuan Z, Dangardt F, Saitz R, Liang W, Chikritzhs T. Selection biases in observational studies affect associations between ‘moderate’ alcohol consumption and mortality. Addiction. 2017 Feb; 112(2):207-14.

Comments received from reviewer: –

I think this is a very reasonable project. This sort of thing has been done before though (John Ioannidis had a paper in JAMA a few years ago comparing trials vs observational studies – I recommend the student look at that)

http://www.ncbi.nlm.nih.gov/pubmed/11497536

My main criticism is that the method for selection of the trials and observational studies is too vague. Why 5 trials? That is completely arbitrary. It would be better to select a defined (sub)area of the subject and then select all observational and all trials in that area.

Then the most logical thing to do is to determine how the outcomes (trials vs Obs) compared as per the Ioannidis paper using methods like meta-analysis.