Residency Research Manual: Your Research Proposal

This section focuses on the writing of your research proposal and what information it should contain. Writing your research proposal provides you with a roadmap for your research project and will be very useful when applying for Institutional Review Board approval for your study.

Your Research Proposal

The previous pages provided you with some starting points as well as thingsto consider when planning your research project. You are now ready to put together your research project proposal.

Answers to the following questions will form the outline for your research project:

• What is my research question?
• Why is it significant?
• What will I measure to support my hypothesis?
• Who will participate?
• How many participants do I need?
• How will I analyze the data?

The answers to questions like these will give you your research project outline and this information needs to be included in your research proposal. Once theproposal is written this process will be much clearer and provide you with a step-by-step approach to finishing your project. Parts of the research design include hypothesis, background, and methods.

Your Question, Your Hypothesis

You already have an idea in mind, the question you want to study and research. Now you need to turn this into a specific hypothesis. A hypothesis is an “educated” guess. They are phrased as statements and are attempts to explain observed patterns or predict the outcome of experiments. Conceptually, your hypothesis is your destination that all research roads must lead to.

A hypothesis is:

• Testable: Variables are operationalized and accessible so it can be tested
• Falsifiable: Should be able to be proven false

The following are examples of both good and bad hypotheses.

Your hypothesis triggers everything you plan to do. Conceptually, think of it as your destination, determining the course of your research and the terminal point of all its pathways.

Choose a hypothesis that is well-focused and testable, and that your experimental results will be able to prove or disprove. Here are examples of EFFECTIVE hypotheses:

From Jacques Banchereau, Ph.D., Baylor University:

• Systemic lupus erythmatosus patients with active disease display considerable alterations in their CD8+ T cell compartments, including effector CD28+ CTLs and suppressor CD28- subsets. We surmise that an excess of killer cells in lupus results in the characteristic tissue damage and explains an excess of dying cells that are considered as key factors in this disease.

From Volker Briken, Ph.D., University of Maryland College Park:

• This proposal seeks to test the hypothesis that the capacity of Mycobacterium tuberculosis to inhibit infection-induced apoptosis of macrophages is a major pathway of the bacteria to avoid the host’s innate and adaptive immune response.

And here are examples of hypotheses that are POORLY focused:

• Understanding the strategies of Escherichia coli to subvert host cells will allow for improved ways of preventing and treating E. coli -related diseases.
• Rheumatoid arthritis patients with active disease show many alterations in their immune profile.
• A wide range of molecules can inhibit HIV infection.

When writing the hypothesis you need to keep in mind the earlier recommendations of FINER. Is this Feasible, Interesting, Novel, Ethical, and Relevant? These are questions that should have been considered earlier but come back at every step and most definitely when you are deciding how to test your hypothesis and how to measure the variables you have identified.

Methods - What and How to Measure

You have your question; you now need to decide what your outcomes are and how they will be measured. If your question is, “What effect does osteopathic manipulation have on the management of knee osteoarthritis?” you need to establish some outcome. This can be a daily pain scale, a quality of life questionnaire, or the amount of medication taken. These are only a couple of examples of possible outcome variables.

Picking a good outcome variable is important and might be determined by what is used in the clinical setting, what you have easy access to, etc. Keep in mind however that precision in the outcome measure and the nature of the outcome measure determines the types of statistical tests you will be able to use and also determine the precision you have in your findings and is linked to how large a sample size you will need.

For example, you can ask the question whether or not there was a change in medication (Yes/No question, a dichotomous variable) or you can look at the amount of medication that was taken (continuous variable). Dichotomous variables provide much less detail and information and will require larger sample sizes to achieve a certain level of statistical power. Continuous outcome variables give more precision and can achieve a greater power with a smaller sample size when compared to dichotomous data.

More detailed information about study design and selection of variables is beyond the scope of this section. There will be plenty of opportunity to delve into this ina more detailed fashion during workshops. However, it is strongly recommended that you use measurement tools (survey, scale, event) which have been well studied and have proven their validity, accuracy, and precision.

Methods - Subject Participation: who, where, and how long?

Determining your sample size is important. The bottom line is that when you have too few participants, your data will lack power and when you have too many participants, your study is inefficient. There is a balance point when after a certain sample size is reached any increase in power no longer outweighs the efforts. Certainly, there are instances, rare conditions and specific chart reviews where you can only take all patients seen during a certain period of time and you may be dealing with very small sample sizes or case studies. The research support team and statistical help on campus can help you with issues related to sample size selection. This is something, which you should do very early on in the design phase of the study.

Who?

Deciding who your study participants are is of great importance. How these will be selected and recruited is important. You can access databases at hospitals, medical offices, or public health agencies with appropriate approval. You canuse patients seen in outpatient clinics, inpatient settings and/or nursing homes. Whichever method you choose to recruit your study participants from has advantages and disadvantages.

The inclusion and exclusion criteria you use for your research project should be similar to published studies. This allows for greater ease of comparing your data.

Your research question also leads you to who the population is that you want to generalize the study findings/results to. This brings you to the question of the clinical/practical significance of the study. Who will be affected by the results? How do the results contribute to our medical understanding or changes in public health policy? The more exclusion criteria, the less generalized the study findings may be.

How Long?

There are several practical issues to consider. One is that your research project is limited by your graduation date. You need to keep a goal in mind of finishing this project in approximately 12 months, start to finish. If you are conducting an experiment this implies that you have 12 months to recruit/obtain participants, conduct the study, collect data, analyze the data and complete the paper. If you are conducting a prospective study and collect data at different time intervals, youalso need to keep in mind this 12 month window of time. This all assumes prior IRB approval, which also can take a couple of months.

If you are joining an ongoing research project, you need to determine if during your 12-month window of time there will be sufficient data for you to analyze and if a midpoint analysis makes sense.

Methods - Data Collection Analysis

The kind of data you collect will determine which analyses you can perform. Unless you have experience in this area, you will want to find someone to assist you. Help should be provided by your mentor but for more difficult studies is available through MWU’s Office of Research and Sponsored Programs. Illinois @ ilorsp@midwestern.edu or Arizona @ azorsp@midwestern.edu

You should have already established a good working relationship with either one of them when deciding on a research topic, selection of sample, etc. You also need a plan as to how and where you will store the collected data and how long you will keep the collected records. It is important to ensure patient confidentiality.

Information should be kept in a locked cabinet in locked offices and a specific plan regarding recordkeeping is needed for your IRB submission. You will also want to design some kind of computerized database for electronic storage of the collected information. This can be done in an excel spreadsheet or in special statistical software. This also is a question to ask your research support group or statistician about early on in the process.

Submitting Your Research Proposal

At this point, you are ready to submit your research proposal for approval. Having completed all the steps described in the previous sections should enable you to complete a 2-5 page research proposal describing your research topic and submit this for approval to your faculty research mentor. This document also needs to be reviewed by your Program Director. This document should include: title, background, hypothesis, specific aims and methods. A research proposaltemplate can be found in Appendix 6. Your research proposal will also serve as the backbone for your application for IRB approval (if needed).

Research Proposal Templates

This first template is the one in use by EM. This is only provided as an example. Please make certain to check if your Program has a specific form which needs to be used.

Name:

Graduating in what year:

Faculty Mentor:

1. State the question or problem your research project is intended to address:
2. What are the underlying assumptions?
3. Provide a relevant literature review.
4. What is your hypothesis?
5. Describe the study population:
6. Describe the method of investigation study:
7. Describe the study protocol:
8. Identify potential limitations in the study design:
9. Describe the approach to data analysis:
10. What is the anticipated application of these results?

This template is taken from the University of Tennessee Research Manual

Project Proposal

Note: each section should be a separate page, double-spaced throughout

1. Title of the project
2. Investigators (your name, mentor's name and any co-investigators). List departmental/institutional affiliations.
3. Abstract brief description of project
4. Specific Aims of Proposed Research. Clearly delineate the intent of the study. Enumerate, in the order of importance, the specific objectives or hypotheses you wish to study.
5. Background and Significance. Briefly sketch the background to your proposal, initially evaluating existing knowledge and specifically identifying the gaps in knowledge which the project is intended to fill. State concisely the importance of the research described in the proposal.
6. Preliminary Studies. This section is a review of work already completed by the investigators. Most of you will leave this section blank.
7. Experimental Design and Methods. Discuss the experimental design and procedures to be used to accomplish the specific aims of the project. Describe the protocol to be used and include the means by which the data will be analyzed and interpreted. Include entry/exclusion criteria and an informed consent form for patients to sign, if applicable. Please note that this should be as brief and succinct as possible, while still providing the information necessary to evaluate the protocol for someone who would not be intimately familiar with the field.

The project proposal plan may be as brief as 2-3 typewritten pages (review paper, case report), or as lengthy as needed. The above format is sufficient for submission to the various research committees, who ultimately grant approval.

© Research Manual for Faculty, Residents, and Medical Students. University of Tennessee College of Medicine Chattanooga.

Research Protocol

This checklist is much more detailed than most in use for the MWU Programs. It is included for your review as it might guide your thinking in developing your research proposal.

Checklist from: ©Manual of basic tools for research in osteopathic manipulative medicine, second edition, April 2007. The University of North Texas Health Science Center Texas College of Osteopathic Medicine.

Guide for Worksheets

The sections in this appendix will help you write statements and draw figures to think though specific parts of your research project. Completing these questions will help you make connections between the parts of the research plan. You will literally sketch out your ideas, steps, and activities you will need to guide you from idea through formulation of hypotheses and design to implementation. Your responses to these items and prompts should take you from the first step on day one to the last step on the day that you will write your report.

Section I: The Question

This section will help you formulate your primary research question or hypothesis. You should finish this section with one sentence. The problem can be stated as a question (e.g. Is there a relationship between Somatic Dysfunction and chronic health problems in persons over age 65 receiving continuing care from a family medicine clinic?) or as a hypothesis (e.g. The worse the chronic health condition a patient has, the more pronounced will be the Somatic Dysfunction. OR if you have two groups you might say Somatic Dysfunction will be more pronounced in persons over age 65 with chronic health problems than in those with no chronic conditions.)

1. Statement of problem of interest:
   Name the condition or disease.
   Name the target population of interest.
   Name the comparison group if any.
2. Is this problem or question of long-term interest like scoliosis, or a does it address conditions related to a temporal event like asthma?
3. Write a brief statement of why the condition of interest is a problem? (This is the “so what” answer. e.g. Restrictions in movement in persons over 65 may exacerbate a chronic health problem leading to overuse of medications for pain or limiting adequate oxygenation). Consider for example whether this is a public health concern, or a cost burden to society.
4. Write one sentence for your main research question or hypothesis. It must withstand the test of being focused and answerable.
5. To answer your question or test your hypothesis what research design will you use? Experimental ____ Number of groups _____, placebo control or standard care control? ______________ Non-experimental ___ e.g. retrospective cohort, cross-sectional prospective, cross-sectional case control, other
6. What exactly will you use to measure your expected results or answer your question?

Section II: Tasks and timelines

1. Right now, as you conceive of it, draw a conceptual flow chart or diagram of exactly what you will do in your research. This helps form the methods section of your proposal. You may come back and change this later, but start now.
2. Right now, as much as you can, associate times such as days and weeks with the tasks needed to achieve your research goals. Start with when it will end and you are writing your report, and work backwards. Include, for example, where you will go, what you will get or see, what tools you will need to collect information and where you will get them, and what your subjects will do and when. Think about who else will be involved from the IRB to your major professor and any staff involved or touched by this project

Section III: Outcomes - Measurements

1. What are the outcome measures of interest?
   These are your variables. Specifically what do you expect to change or to happen as a result of the intervention? Or what are you going to analyze as a measure of change or difference? How will you define your outcomes of interest? i.e. in the example used say: For this study Somatic Dysfunction will be the total number of lesions documented in the chart at the time of the first visit. For a nonexperimental cross sectional descriptive study, consider for example what you want to examine such as asking if a certain group of patients show medical or health status changes following OMT for a particular condition such as back pain?
2. How will you know the answer to your question? _______________________________________________ _________________________________________________________________________________
3. What information will you collect?
   How will you measure each of your variables? Think about how you will be consistent especially if different source documents use different measurements (i.e. inches compare to inches, palpatory findings must be consistent or you must consider how many different practitioners’ measurements you will use and consider this in the power analysis)
   Define the outcome: i.e. don’t just say blood pressure, say health status will be measured by the following criteria. 1. sputum will be collected as …… x times….. and tested using….. the values to be included in this study are ……, 2. blood pressure will be recorded at five minutes after the rest period following collection of heart rate…….

   Outcome variable	Definition	Source

4. If you are using health indicators are they in a record or do you have to collect them? If you are asking questions about functionality is there an existing measurement instrument or questionnaire or do you have to create one?
   List the data collection forms (DCFs).

   DCF	Source and other information

5. Name the Independent (what will not vary or change) and Dependent Variables (your outcome variables of interest measured as change in a condition, a number of actions or events, or a status)
6. Are you comparing change in different age groups with a particular disease who used an OMT clinic or those who got OMT and those who did not, for a particular condition? List any grouping variables that will be independent variables or factors such as age, gender, presence or absence of another condition or quality.
7. List the variables you will test as predictor or criterion variables, such as a number of prenatal visits could predict birth weight of baby for women who had OMT

Section IV: Inclusion/Exclusion Criteria

1. List the characteristics or criteria by which you will select your records or subject for the study (inclusion criteria). List those by which you will exclude records or subjects (exclusion criteria). (e.g. age, health condition, lab value, time periods, services criteria, locations)

Section V: Acquisition and Randomization

1. How will you acquire the sample of records or persons? Where will you get the information or subjects, how will you recruit or access them?
2. If this is an experimental design, what are the experimental groups? How will you make random assignment?

Section VI: Review and Revise

1. REVIEW: Look back at the question or hypothesis you wrote on page one of the worksheets. REWRITE one sentence for your main research question or hypothesis. It must withstand the test of being focused and answerable
2. Ask someone to look at the question you wrote and ask you questions about it. Have a discussion to refine these key items.
3. After revising or re-writing your primary question or hypothesis, review the drawing of your research question, methods and outcomes and the time-table in items 1 and 2. REVISE THESE IN THE SPACE BELOW IF YOU NEED TO DO THAT NOW.
   1. Flow chart or diagram of exactly what you will do in your research.
   2. Time-table

      What/Task	When	Who
      Power analysis to determine sample size using information from published research on effect size – amount of change or difference anticipated	When I have my question or hypothesis finished	A biostatistician

Section VII: Data Management and Analysis

1. REVIEW AND REVISE THE DCF LIST AND USE IT TO ANSWER THESE QUESTIONS
2. What will your data collection forms look like? Attach samples. If you develop your own you must test them and adjust them before you begin the data collection.
   1. Are the collection instruments in the public domain, or is a cost associated? Is a license required? (It is important to not accept someone’s opinion that you can use an instrument. You would want credit given to you if you developed an instrument. Also, there are many good public domain or other-student-developed instruments available.)
   2. How will you document validity and/or reliability of the measurement instruments?
   3. Will you attempt to “generalize” to a larger population or other populations?
   4. Which computer software will you use to enter data?
   5. How and where will data be stored?
   6. How will you protect the data from errors or tampering?
   7. How will data be guaranteed as confidential or anonymous?
   8. Identify specific data analysis approach and appropriate statistical tests of the data.
   9. Identify data characteristics that may influence the analysis methods.

Checklist from: ©Manual of basic tools for research in osteopathic manipulative medicine, second edition, April 2007. The University of North Texas Health Science Center Texas College of Osteopathic Medicine