This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License, which allows others to remix, tweak, and build upon the work non-commercially, as long as the author is credited and the new creations are licensed under the identical terms.
This concluding module attempts to synthesize the key learning points discussed during the course of the previous ten sets of modules on methodology and biostatistics. The objective of this module is to discuss how to present a model research proposal, based on whatever was discussed in the preceding modules. The lynchpin of a research proposal is the protocol, and the key component of a protocol is the study design. However, one must not neglect the other areas, be it the project summary through which one catches the eyes of the reviewer of the proposal, or the background and the literature review, or the aims and objectives of the study. Two critical areas in the “methods” section that cannot be emphasized more are the sampling strategy and a formal estimation of sample size. Without a legitimate sample size, none of the conclusions based on the statistical analysis would be valid. Finally, the ethical parameters of the study should be well understood by the researchers, and that should get reflected in the proposal.
K EY W ORDS : Biostatistics, research proposal, research protocol, study designAs we reach the end of an exhaustive module encompassing research methods and biostatistics, we need to summarize and synthesize the key learning points, to demonstrate how one may utilize the different sections of the module to undertake research projects of different kinds. After all, the practical purpose behind publishing such a module is to facilitate the preparation of high quality research proposals and protocols. This concluding part will make an attempt to provide a window to the different sections of the module, underlining the various aspects of design and analysis needed to formulate protocols applicable to different kinds of clinical research in dermatology.
The goal of a research proposal is to present and justify the need to study a research problem and to present the practical ways in which the proposed study should be conducted. A research proposal is generally meant to be presented by an investigator to request an agency or a body to support research work in the form of grants. The vast majority of research proposals, in India, however, are not submitted to agency or body for grants, simply because of the paucity of such agencies, bodies, and research grants. Most are academic research proposals, self-financed, and submitted to scientific and ethics committee of an institution. The parts of a proposal include the title page, abstract/project summary, table of contents, introduction, background and review of literature, and the research protocol.
The title page should contain the personal data pertaining to the investigators, and title of the project, which should be concise and comprehensive at the same time. The table of contents, strictly speaking, is not necessary for short proposals. The introduction includes a statement of the problem, purpose, and significance of the research.
The protocol is the document that specifies the research plan. It is the single most important quality control tool for all aspects of a clinical research. It is the instrument where the researcher explains how data will be collected, including the calculation for estimating sample size, and what outcome variables to measure.
A complete clinical research protocol includes the following:
Study design Precise definition of the disease or problemCompletely defined prespecified primary and secondary outcome measures, including how and when these will be assessed
Clear description of variables Well-defined inclusion and exclusion criteria Efficacy and safety parameters Whenever applicable, stopping guidelines and parameters of interim analyses Sample size calculation Randomization details Plan of statistical analysis Detailed description of interventions A chronogram of research flow (Gantt chart) Informed consent document Clinical research form Details of budget; and References.(Modified from: Bagatin et al., 2013).
The project summary is a brief document that consists of an overview, and discusses the intellectual merits, and broader impacts of the research project. Each of these three sections is required to be present and must be clearly defined. The project summary is one of the most important parts of the proposal. It is likely the first thing a reviewer will read, and is the investigators’ best chance to grab their interest, and convince them of the importance, and quality, of their research before they even read the proposal. Though it is the first proposal element in order, many applicants prefer to write the project summary last, after writing the protocol. This allows the writer to better avoid any inconsistencies between the two.
The overview specifies the research goal and it should demonstrate that this goal fits with the principal investigator's long-term research goals. It should specify the proposed research approach and the educational goal of the research project.
The intellectual merits (the contribution your research will make to your field) should specify the current state of knowledge in the field, and where it is headed. It should also clarify what your research will add to the state of knowledge in the field. Furthermore, important to state is what your research will do to enhance or enable other researches in the field. Finally, one should answer why your research is important for the advancement of the field.
The broader impacts (the contribution the research will make to the society) should answer the questions on the benefit to the society at large from the research, and the possible applications of the research, and why the general public would care. It should also clarify how the research can benefit the site of research (medical college or university, etc.) and the funding agency.
This is an important component of the research protocol. The review should discuss all the relevant literature, the method used in the literature, the lacunae in the literature, and justify the proposed research. We have provided a list of the useful databases in the section on systematic reviews and meta-analysis (Setia, 2017). Some of these are PubMed, Cochrane database, EMBASE, and LILACS.
The researcher should not provide a descriptive analysis of literature. For instance, the literature reviews should not be a list of one article followed by the next article. It should be a critical analysis of literature.
A study by XXXX et al. found that the prevalence of psoriasis was 20%. It was a hospital-based study conducted in North India. The prevalence was 35% in males and 12% in females.
Another study by YYYYY et al. found that the prevalence of psoriasis was 14%. The study was conducted in a private clinic in North India. The prevalence was 8% in males and 18% in females.
A third study by ZZZZZ et al. found that the prevalence of psoriasis was 5%. This study was a community-based study. The prevalence was 7% in males and 3% in females.
In this type of review, the researcher has described all the studies. However, it is useful to understand the findings of these three studies and summarize them in researcher's own words.
A possible option can be “The reported prevalence of psoriasis in the Indian population varied from 5% to 20%. In general, it was higher in hospital-based studies and lower in community-based studies. There was no consistent pattern in the prevalence of psoriasis in males and females. Though some studies found the prevalence to be higher in males, others reported that females had a higher prevalence.”
The researcher should discuss the limitations of the studies. These could be the limitations that the authors have presented in the manuscript or the ones that the researcher has identified. Usually, the current research proposal should try to address the limitations of a previous study.
A study by BBBB et al: “One of the main limitations of our study was the lack of objective criteria for assessing anemia in patients presenting with psoriasis. We classified the patients based on clinical assessment of pallor.”
The present proposal can mention “Though previous studies have assessed the association between anemia and psoriasis, they have not used any objective criteria (such as hemoglobin or serum ferritin levels). Furthermore, pallor was evaluated by three clinicians; the authors have not described the agreement between these clinicians.”
In the above example, the authors have stated the limitation of their research in the manuscript. However, in the review of literature, the researcher has added another limitation. It is important to convince the reviewers that the researcher has read and understood the literature. It is also important that some or most of these lacunae should be addressed in the present proposal as far as possible.
The researcher should adequately justify the present proposal based on the review of literature. The justification should not only be for the research question, but also the methods, study design, variables of interest, study instruments or measurements, and statistical methods of choice. Sometimes, the justification can be purely statistical. For example, all the previous studies have used cross-sectional data or cross-sectional analysis of longitudinal data in their manuscripts. The present proposal will use methods used for longitudinal data analysis. The researcher should justify the benefit of these methods over the previous statistical methods.
In short, the review should not be a “laundry list” of all the articles. The review should be able to convince the reader that the present research is required and it builds on the existing literature (either as a novel research question, new measurement of the outcome, a better study design, or advanced and appropriate statistical methods).
Kindly try to avoid this justification: “It has not been done in our center.”
The “aim” of the study is an overarching goal of the study. The objectives are measurable and help the researcher achieve the overall aim.
For example, the overall aim of our study is to assess the long-term health of patients of psoriasis.
The specific objectives are:
To record the changes in Psoriasis Area and Severity Index (PASI) score in patients with psoriasis over a period of 5 years
To study the side effects of medications in these patients over a period of 5 years.It is important to clearly state the objectives, since the research proposal should be designed to achieve these objectives.
For example, the methods should describe the following:
How will the researcher answer the first objective? Where will the researcher recruit the study participants (study site and population)? Which patients of psoriasis will be recruited (inclusion and exclusion criteria)? What will be the design of the study (cohort, etc.)?What are all the variables to be measured to achieve the study outcomes (exposure and outcome variables)?
How will the researcher measure these variables (clinical evaluation, history, serological examination, etc.)?
How will the researcher record these data (clinical forms, etc.)? How will the researcher analyze the data that have been collected?Are there any limitations of these methods? If so, what has the researcher done to minimize the limitations?
All the ten modules on research methodology have to be read and grasped to plan and design any kind of research applicable to one's chosen field. However, some key areas have been outlined below with examples to appreciate the same in an easier manner.
The study setting must be specified. This should include both the geographical location and the population from which the study sample would be recruited.
“The study took place at the antiretroviral therapy clinic of Queen Elizabeth Central Hospital in Blantyre, Malawi, from January 2006 to April 2007. Blantyre is the major commercial city of Malawi, with a population of 1,000,000 and an estimated HIV prevalence of 27% in adults in 2004” (Ndekha et al., 2009).
This is a perfect example of description of a study setting which underscores the importance of planning it in detail a priori.
Study population has to be clearly and precisely defined. For example, a study on atopic dermatitis may be conducted upon patients defined according to the UK Working Party's modified diagnostic criteria, or the Hanifin and Rajka's criteria, or some other criteria defined by the investigators. However, it should always be prespecified within the protocol.
Similarly, the eligibility criteria of the participants for the study must be explicit. One truism that is frequently forgotten is that the inclusion and exclusion criteria are mutually exclusive, and one is not the negative image of the other. Eligible cases are included according to a set of inclusion criteria, and this is followed by administration of the exclusion criteria. Thus, in fact, they can never be the negative image of each other.
“Eligible participants were all adults aged 18 or over with HIV who met the eligibility criteria for antiretroviral therapy according to the Malawian national HIV treatment guidelines (WHO clinical stage III or IV or any WHO stage with a CD4 count < 250/mm 3 ) and who were starting treatment with a BMI < 18.5. Exclusion criteria were pregnancy and lactation or participation in another supplementary feeding program” (Ndekha et al., 2009).
To put in perspective the point we made about inclusion and exclusion criteria, in the above example, “age above 18 years” or “CD4 count >250/mm 3 ” cannot be exclusion criteria, as these have already been excluded.
Sampling strategy has been adequately discussed in the Module 5 of the Methodology series (Setia, 2016). A few points are worth repeating:
The sampling strategy should never be misrepresented. Example: If you have not done random sampling, no big deal. There are other legitimate sampling strategies available for your study. But once you have mentioned “random sampling” in your protocol, you cannot resort to purposive sampling
Sometimes, the researcher might want to know the characteristics of a certain problem within a specific population, without caring for generalizability of results. In such a scenario, purposive sampling may be resorted to
Nonprobability sampling methods such as consecutive consenting sampling or any such convenience sampling are perfectly legitimate and easy to do, particularly in case of dissertations where time and resources are limited.
Sample size is one of the most misunderstood, yet fundamentally important, issues among clinicians and has to be addressed once the study objectives have been set and the design has been finalized. Too small a sample means that there would be a failure to detect change following test intervention. A sample larger than necessary may also result in bad quality data. In either case, there would be ethical problems and wastage of resources. The researcher needs just enough samples to draw accurate inferences, which would be adequately powered (Panda, 2015).
Estimation of sample size has been dealt with adequately in the Module 5 biostatistics series (Hazra et al., 2016), including the different mathematical derivations and the available software. Sample size determination is a statistical exercise based on the probability of errors in testing of hypothesis, power of the sample, and effect size. Although, relatively speaking, these are simple concepts to grasp, a large number of different study designs and analytical methods lead to a bewilderingly large number of formulae for determining sample size. Thus, the software are really handy and are becoming increasingly popular.
The study design defines the objectives and end points of the study, the type and manner of data collection, and the strategy of data analysis (Panda 2015). The different types of clinical studies have been depicted in Figure 1 . The suitability of various study designs vis-à-vis different types of research questions is summarized in Table 1 .
Types of study (Source: Panda, 2015)
Research questions vis-a-vis study designs
In our previous series of ten modules on methodology, we have discussed all these different kinds of studies and more. Some key issues that require reiteration are given below:
The control of a case–control study and that of a randomized controlled trial is more different from each other than chalk is from cheese. The former is an observational study, while the latter is an interventional one. Every study with a control group is not a case–control study. For a study to be classified as a case–control study, the study should be an observational study and the participants should be recruited based on their outcome status (Setia, 2016). Apparently, this is not so difficult to understand, yet even now we have publications which confuse between the different kinds of controls (Bhanja et al., 2015)
Due to the fact that the outcome and exposure are assessed at the same time point in a cross-sectional study, it is pretty difficult, if not impossible, to derive causal relationships from such a study. At most, one may establish statistical association between exposures and outcomes by calculating the odds ratio. However, these associations must not be confused with causation.
It is generally said that a cohort design may not be efficient for rare outcomes. However, if the rare outcome is common in some exposures, it may be useful to follow a cohort design. For example, melanoma is a rare condition in India. Hence, if we follow individuals to study the incidence of melanoma, it may not be efficient. However, if we know that, in India, acral lentiginous melanoma is the most commonly reported variant, we should follow a cohort of individuals with acral lentiginous and study the incidence of melanoma in this group (Setia, 2016).
Clinical researchers should also be accustomed with observational designs beyond case–control, cohort, and cross-sectional studies. Sometimes, the unit of analysis has to be a group or aggregate rather than the individual. Consider the following example:
The government introduced the supplementation of salt with iodine for about 20 years. However, not all states have used the same level of iodine in salt. Certain hilly states have used higher quantities compared with other states. Incidentally, you read a report that high iodine levels are associated with psoriasis. You are intrigued to find if introduction of iodine has altered the picture of psoriasis in the country. You feel compelled to design a study to answer this question.
It is obvious that here the unit of study cannot be individuals, but a large population distributed in a certain geographical area. This is the domain of ecologic studies. An allied category of observational studies is named “natural experiments,” where the exposure is not assigned by the investigator (as in an interventional study), but through “natural processes.” These may be through changes in the existing regulations or public policies or, may be, through introduction of new laws (Setia, 2017).
Another category of research questions that cannot be satisfactorily captured by all the quantitative methods described earlier, like social stigma experienced by patients or their families with, say, vitiligo, leprosy, or sexually transmitted infections, are best dealt with by qualitative research. As can be seen by the examples given above, this is a type of research which is very relevant to medical research, yet to which the regular medical researcher has got a very poor exposure, if any. We shall encourage interested researchers to take a look at the 10 th Module of the Methodology series that specifically deals with qualitative research (Setia, 2017).
Clinical studies are experiments that are not conducted in laboratories but in controlled real-life settings on human subjects with some disease. Hence, designing a study involves many pragmatic considerations aside pure methodology. Thus, factors to consider when selecting a study design are objectives of the study, time frame, treatment duration, carryover effects, cost and logistics, patient convenience, statistical considerations, sample size, etc. (Panda, 2015).
Certain truisms regarding study designs should always be remembered: a study design has to be tailored to objectives. The same question may be answered by different designs. The optimum design has to be based on workforce, budgetary allocation, infrastructure, and clinical material that may be commanded by the researchers. Finally, no design is perfect, and there is no design to provide a perfect answer to all research questions relevant to a particular problem (Panda, 2015).
Data structure depends on the characteristics of the variables [ Figure 2 ]. A variable refers to a particular character on which a set of data are recorded. Data are thus the values of a variable (Hazra et al., 2016).
