The formulation of a specific question is the most important concrete step in the development of a research project. After the question is conceived it is further clarified by a review of the appropriate literature. This is often not a linear process - the literature review often engenders its own questions and substantially modifies the initial idea. Regardless of when an idea is approaching its final form, the researcher must keep the following three items in mind:
- the population,
- the intervention, and
- the outcome. (Research questions are more likely to be published if they are thought provoking regardless of whether proven or disproved.)
Has the question already been answered? What hypothesis or expected result does the literature support? Resources that may be useful in a review of the pertinent literature include:
Medline
Abstracts of the medical literature:
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi
http://www.sciencekomm.com/
Cochrane Collaboration:
Collection of rigorous systematic reviews:
http://www.cochrane.org/
Full text articles for many journals are available
Follow library links at:
Columbia University Health Sciences Library:
http://library.cpmc.columbia.edu/hsl/eres/digital.cfm
use the link on the upper toolbar, second from left, "digital collections" subheading "E-journals" subheading "all E-journals" - the other links on the page give only partial access.
Cornell University Weill Medical College Library:
http://lib.med.cornell.edu/cgi-bin/EJournal.cgi
Science citation index: allows searching forward and backward in time based on references "this work was referenced by…" or " this work references…..": http://www.columbia.edu/cu/libraries/indexes/science-cit-ind.html
Significance
Does this study address an important problem? Also known as the "so what?" criteria. The answer to this question is often best obtained by reviewing the literature and by talking with experienced colleagues.
Approach
Are the design, methods and analyses well developed and appropriate? For example, is the sample size feasible for the resources available? We can arrange a formal statistical consult through the Department of Biostatistics, or find an appropriate statistician to consult informally on Emergency Medicine research projects.
Environment and available resources
Does the scientific environment in which the work will be done contribute to the probability of success? For example, can you capture and follow-up the needed number of patients in a reasonable amount of time and with the resources on hand, or will you require outside funding or additional manpower?
Statistics
Answering questions about available resources requires an estimate of the number of subjects/patients needed to to make meaningful conclusions about the results. This requires "power analysis", a set of calculations to determine how many participants might be required to show a statistically significant result given a baseline variability of outcome and a given size of treatment effect.
One web based tutorial available for this purpose is "G*Power" available at:
www.psycho.uni-duesseldorf.de/aap/projects
/gpower/how_to_use_gpower.html
Another useful statistics page, or rather a posting of MANY statistics pages, is:
http://members.aol.com/johnp71/javastat.html
Two by Two Table analysis including calculation of sensitivity, specificity, likelihood ratios:
http://www.hutchon.freeserve.co.uk/EPRval.htm
Critical appraisal checklist and calculator for an article on therapy or prevention. This calculator works off-line:
http://www.hutchon.freeserve.co.uk/Single-trial.htm
Critical Appraisal - diagnostic test. This calculator works off-line:
http://www.hutchon.freeserve.co.uk/Diagnostic-test.htm
Calculator for confidence intervals of odds ratio in an unmatched case control study. For example, groups of cases and controls studied to assess a treatment or exposure to a suspected causal factor. This calculator works off-line:
http://www.hutchon.freeserve.co.uk/ConfidOR.htm
Calculator for converting risk to odds and vice versa This calculator works off-line:
http://www.hutchon.freeserve.co.uk/Oddstorisk.htm
Calculator for confidence intervals of odds ratio in an unmatched case control study. For example groups of cases and controls studied to assess exposure to a suspected causal factor. This calculator works off-line:
http://www.hutchon.freeserve.co.uk/ConfidORnulhypo.htm
Calculator for survival probabilities (the Kaplan-Meier method)
20 years (other time interval such as month, etc can be substituted) This calculator works off-line:
http://www.hutchon.freeserve.co.uk/Kaplan-Meier.htm
Web calculator for use with an article on risk. This calculator works off-line:
http://www.hutchon.freeserve.co.uk/harm.htm
Testing the difference between means. Two samples taken from within a normally distributed population. This calculator works off-line:
http://www.hutchon.freeserve.co.uk/Diffbetwmeans.htm
Confidence intervals for proportions, using Wilson's method. This calculator works off-line:
http://www.hutchon.freeserve.co.uk/Wilsons.htm
Calculator for confidence intervals of relative risk:
http://www.hutchon.freeserve.co.uk/ConfidRR.htm
VassarStats web page with multiple calculators:
http://faculty.vassar.edu/lowry/VassarStats.htm
Getting Funded:
It may be that there is little chance of a project's success without financial support. There are many sources of funding available. Residents will be encouraged to find outside sources of funding, and be supported in their grant applications by appropriate members of the faculty.
Grantsnet: free grant search program:
http://www.grantsnet.com/
The NIH:
The main source of government directed grant funding in the USA is the National Institutes of Health (NIH):
http://www.nih.gov
Columbia and Cornell grants
Columbia University's Health Sciences Center Faculty and Research Information Homepage has links to various funding sources:
http://cpmcnet.columbia.edu/research/
Columbia University's Health Sciences Center Office of Clinical Trials - information on ongoing research with private sector sponsorship:
http://cpmcnet.columbia.edu/dept/ctrials/
Cornell University's Weill Medical College Grants and Contracts Home Page:
http://www.med.cornell.edu/research/grants2/index.html
The Foundation Center - links to private foundations that fund research:
http://fdncenter.org/grantmaker
Submitting a proposal:
Projects must be submitted to an institution's Institutional Review Board (IRB), which ensures patient safety and ethical research. The Board will decide if proposals are eligible for "full review" or "expedited review". For example, if a study only involves manipulation of data that has been stripped of patient identifiers, it may be eligible for exemption from review.
Forms for IRB submission are available at the IRB websites for Columbia University Health Sciences Center:
http://www.cpmcirb.hs.columbia.edu/root/main.htm
For Weill Medical College of Cornell University (through intranet only), forms are available at:
http://www.med.cornell.edu/research/offices/index.html
Writing the paper
Using a carefully planned approach to paper writing will save tremendous time in the end. If you have produced a well thought out proposal, you have already accomplished much of the work.
Brief essay on writing a scientific paper:
http://aerg.canberra.edu.au/pub/aerg/edupaper.htm
The following outline may prove useful:
- Introduction / Abstract
- Objectives and Rationale
- Study Design
- Description of Study Population
- Inclusion / Exclusion Criteria
- Subject Recruitment Methods
- Study Management /Personnel
- Methods and Procedures
- Drug / Device Information
- Safety Profile
- Potential Risks to Subjects
- Data Handling and Confidentiality
- Baseline Measurements
- Outcome Measurements
- Post-Study Measurements
- Data Analysis and Evaluation Technique
- Limitations
- Bibliography
Once a paper is in the process of being written, one might consider presenting it in abstract form at a national or regional meeting:
ACEP
http://www.acep.org/calendar/index.cfm?type=1&keyword=9
SAEM
http://www.saem.org/inform/calendar.htm
Instructions for authors
The final written paper must be formatted appropriately for the Journal to which it will be submitted. The following web page, from the Medical College of Ohio, links to instructions for authors pages of hundreds of scientific journals: http://www.mco.edu/lib/instr/libinsta.html
Academic Resources
