Dr Somesh Ojha, BDS · Clinical research methodology & biostatistics
Research designed to hold up under scrutiny.
Study design, statistical analysis and scientific writing for clinical researchers, postgraduate and doctoral candidates, and publication teams. I work from topic selection and protocol through analysis to a submitted manuscript or a submitted thesis — with the methods documented well enough to defend at review, or at viva.
- Practising since 2022
- Bagdogra, India
- Remote engagements
Services
The full arc, or any single stage of it
Some projects arrive as a blank page and a supervisor's deadline. Others arrive as a dataset nobody has analysed, or a manuscript that has come back from review. I take projects at any point along that line, and I am equally willing to do one stage well as to run the whole thing.
Design and protocol
-
Topic selection and title formulation
Scoping the existing literature to find a question that is answerable with the resources you actually have, has not already been settled, and will survive an institutional review committee. Framed in PICO or PECO terms, with several defensible title options rather than one.
-
Protocol and synopsis development
Complete protocols and synopses for institutional and ethics committee submission: background, aims and objectives, methodology, sampling strategy, sample size with justification and power calculation, statistical analysis plan, timeline, references and annexures.
-
Questionnaire and instrument development
Item generation from the literature, expert panel review, content validity indexing, face validity, pilot testing, internal consistency and test–retest reliability, and exploratory or confirmatory factor analysis where the instrument warrants it.
-
Protocol registration and institutional documentation
PROSPERO records for systematic reviews, CTRI for clinical trials, and the supporting paperwork institutional review boards and ethics committees ask for — participant information sheets, consent forms, case record forms and data collection proformas. Registration is covered in detail below.
Analysis and synthesis
-
Original research analysis
Trials and observational studies alike: data cleaning and recoding, the analysis plan, descriptive and inferential testing, regression and survival modelling, agreement and reliability statistics, and a results section written to match the tables rather than to describe them.
-
Systematic review and meta-analysis
Search strategy design across PubMed, Scopus, Embase and the Cochrane Library, dual screening, structured extraction, random- or fixed-effects pooling, and full heterogeneity, subgroup and sensitivity analysis.
-
Narrative, scoping and umbrella reviews
Structured narrative synthesis where pooling is not defensible, scoping reviews to PRISMA-ScR, and reviews of existing reviews — with the same discipline applied to searching and selection that a meta-analysis would get.
-
Figures and data visualisation
Forest and funnel plots, PRISMA flow diagrams, Kaplan–Meier and ROC curves, risk-of-bias summaries, GRADE summary-of-findings tables, graphical abstracts and composite figures — built at 300 DPI to the target journal's specification.
Writing and submission
-
Thesis and dissertation support
Chapter structuring, review of literature, methodology and results writing, statistical analysis, formatting to university guidelines, reference management, and preparation to defend the methods at viva. Detailed in the next section.
-
Manuscript preparation
Full drafting and restructuring to IMRaD, thesis-to-manuscript conversion, journal-specific reformatting, Vancouver or APA reference handling, and cover letters written to the editor rather than to nobody.
-
Revision and resubmission
Point-by-point responses to reviewers, the additional analyses reviewers ask for, rebuttals where the criticism is answerable, and reformatting for a new target journal after rejection.
Study designs
Designs worked across
The design determines the statistics, the reporting checklist and half the reviewer's questions. These are the families I work in and what support looks like in each.
- Randomised controlled trialsParallel-group, crossover and split-mouth designs; randomisation sequence generation, allocation concealment, blinding strategy, intention-to-treat handling, CONSORT reporting and SPIRIT-compliant protocols.
- Cohort studiesProspective and retrospective; exposure definition, loss to follow-up, time-to-event and survival analysis, confounder adjustment, STROBE reporting.
- Case–control studiesMatching strategy, odds ratio estimation, conditional and unconditional logistic regression, confounder control.
- Cross-sectional and KAP surveysSampling frame and sample size, instrument validation, prevalence estimation with confidence intervals, association testing and multivariable modelling.
- Diagnostic and prognostic accuracySensitivity, specificity, predictive values and likelihood ratios, ROC and AUC analysis, model calibration, STARD and TRIPOD reporting.
- In vitro and laboratory studiesGroup allocation and replicate handling, ANOVA-family designs with appropriate post-hoc comparison, and figures built for a materials or basic-science readership.
- Case reports and case seriesCARE-compliant structure for cases with genuine teaching value, including timeline figures and clinical image handling.
- Evidence synthesisSystematic, scoping, umbrella and narrative reviews, with meta-analysis where pooling is statistically and clinically defensible — and an honest narrative where it is not.
Protocol registration
PROSPERO, CTRI and the alternatives
Prospective registration is what separates a planned review from one reconstructed after the fact, and most journals now want the number before a manuscript goes out to reviewers. It is also where a great many postgraduate projects stall — the record comes back refused, and nobody explains why.
- PROSPEROThe international prospective register of systematic reviews, maintained by the Centre for Reviews and Dissemination at the University of York, and open to systematic reviews with a health-related outcome. I draft the full record, answer CRD's queries, resubmit where a refused review is still eligible, and log amendments as the protocol changes during the project.
- CTRIThe Clinical Trials Registry – India. Prospective registration is mandatory for trials conducted in India and is checked by both ethics committees and journals. I prepare the submission alongside the protocol and supporting documentation.
- Open Science FrameworkWhere PROSPERO does not apply — scoping reviews, methodological reviews, original study protocols — OSF provides a timestamped, citable public record that serves the same purpose for reviewers.
Why records get refused. Four reasons account for most of them: data extraction is already under way, so the review is no longer prospective; the review has no health-related outcome; it is a scoping review, which PROSPERO does not accept; or the protocol is too thin to assess — eligibility criteria, search strategy and analysis plan all have to be specified at submission, not afterwards. Getting these right first time is the difference between a registration number and several lost weeks.
Academic and institutional work
Dissertation, thesis and departmental support
A large share of my work is with postgraduate and doctoral candidates — MDS, MD, MS, MSc, DNB, MPhil and PhD — and with the departments and research cells that supervise them. The pressure points are consistent: a topic that will not clear the committee, a protocol due next week, a dataset that arrived without an analysis plan, and a viva where the methods have to be explained out loud.
- Topic and title
- Feasibility scoping, gap identification, and title options framed to what your department, sample access and timeline can actually support
- Synopsis and protocol
- Full synopsis and protocol drafted to your university's prescribed format, ready for institutional review board and ethics committee submission
- Review of literature
- Comprehensive chapter-length reviews, study by study, with statistics carried through the narrative and citations in the style your university prescribes
- Methodology
- Sampling, sample size and power justification, armamentarium, procedure description, outcome definitions and the statistical analysis plan
- Statistics and results
- Analysis of your collected data, master chart preparation, tables and figures to specification, and the results chapter written around them
- Discussion and conclusion
- Interpretation set against the existing literature, limitations stated honestly, and conclusions that the data will actually bear
- Formatting and submission
- University template compliance, reference management, annexures, and preparation for plagiarism screening
- Viva preparation
- A walk-through of your own methods and statistics so the choices are yours to defend, not mine to explain
- Departmental work
- Research cell support: proposal development, funding and grant applications, institutional projects, and methodology guidance for teams running several studies at once
On scope. Every university sets its own regulations on what assistance a candidate may receive, and those regulations — not my brief — decide where the line sits on any given thesis. I work within what your guide and institution permit, put the scope in writing before starting, and will tell you plainly when a request falls outside it. The methods stay yours to defend.
Methods
The methodological inventory
A study is only as defensible as the choices behind it, and those choices are fixed in the protocol rather than discovered in the data. These are the instruments and models I work with.
- Reporting standards
- CONSORT, SPIRIT, STROBE, STARD, TRIPOD, PRISMA 2020, PRISMA-ScR, MOOSE, CARE, COREQ, SQUIRE
- Registration
- CTRI, PROSPERO, Open Science Framework
- Design and sampling
- Sample size and power calculation, randomisation sequence generation, allocation concealment, blinding, crossover and split-mouth designs, sampling frame construction
- Instrument development
- Item generation, content validity index, face validity, pilot testing, Cronbach's alpha, test–retest reliability, exploratory and confirmatory factor analysis
- Inferential statistics
- t-tests, ANOVA, ANCOVA and repeated-measures designs, non-parametric equivalents, chi-square and Fisher's exact, correlation, linear, logistic and Cox regression, survival analysis, ROC and diagnostic accuracy
- Agreement and reliability
- Cohen's and Fleiss' kappa, intraclass correlation, Bland–Altman analysis
- Synthesis models
- DerSimonian–Laird, REML, Mantel–Haenszel, inverse variance, Freeman–Tukey double arcsine for proportions
- Heterogeneity
- Cochran's Q, I², τ², prediction intervals, leave-one-out and subgroup analysis, meta-regression
- Risk of bias
- RoB 2, ROBINS-I, Newcastle–Ottawa Scale, JBI critical appraisal checklists, QUADAS-2
- Certainty of evidence
- GRADE, GRADEpro summary-of-findings tables
- Software
- SPSS, R, RevMan, GRADEpro, G*Power, Python for figure production, EndNote and Zotero for reference management
Clinical domains worked in
Selected work
Projects, described at method level
Client work is confidential, so these are summarised by design and analysis rather than by title, candidate, institution or outcome.
- Split-mouth RCT
Microneedling and ascorbic acid in gingival depigmentation
Randomised split-mouth trial; allocation and blinding scheme, repigmentation scoring, repeated-measures analysis across time points.
- Randomised trial
Behaviour guidance with a dental simulation game
Paediatric RCT comparing Tell–Show–Do with and without a digital adjunct; pain and anxiety outcomes, CONSORT-compliant reporting.
- Retrospective cohort
Fluid balance after neonatal cardiac surgery
Results section built from a 151-column clinical database, including derived variables, outcome modelling and figure production.
- Retrospective study
CBCT in dental implant treatment planning
Record-based analysis of diagnostic and planning parameters; descriptive and comparative statistics, tables and figures.
- Laboratory study
Zirconia surface conditioning and bond strength
In vitro comparison across conditioning protocols; group allocation, ANOVA with post-hoc comparison, thesis-to-manuscript conversion.
- Experimental model
Photobiomodulation and angiogenesis in the CAM model
Single- versus dual-mode diode laser irradiation in the chick chorioallantoic membrane; vessel quantification and analysis.
- Cross-sectional
Atopic dermatitis burden and environmental exposure
Cross-sectional and ecological analyses; statistical work identifying demographic confounding, with publication-quality figures.
- Meta-analysis
Intravenous versus oral iron in pregnancy
Pooled maternal haematological and neonatal outcomes with a full forest plot suite and evidence tables.
- Meta-analysis
Topical retinoids in congenital ichthyosis
Single-arm proportion meta-analysis using the Freeman–Tukey double arcsine transformation.
- Systematic review
Virtual versus in-person pulmonary rehabilitation in COPD
Six forest plots, seven evidence tables and a full risk-of-bias summary prepared for submission.
- Diagnostic models
Machine-learning models in myopia
Accuracy synthesis across heterogeneous reporting, with GRADEpro certainty tables.
- Dissertations
Postgraduate theses across specialities
Full dissertations in paediatric dentistry, prosthodontics and oral medicine — literature review, methodology, statistics and formatting to university template.
Anonymised deliverables — forest plots, evidence tables, risk-of-bias figures, GRADE tables and statistical output — can be shared on request where the client has approved their release.
Process
How a project runs
Sequential, and deliberately so — the analysis plan is fixed before the first observation is recorded or the first abstract screened, because a protocol written after the results are known is not a protocol.
- 01
Scoping
The question in PICO terms, a scan of what already exists, and an honest read on whether the study is feasible with your sample, timeline and resources.
- 02
Protocol
Design, eligibility, sampling, sample size, outcome definitions and the statistical analysis plan — written up for committee submission and registered on PROSPERO, CTRI or OSF as the design requires.
- 03
Instruments
Proformas, case record forms, questionnaires and validation work, or the search strategy and screening plan if the project is a review.
- 04
Data
Master chart design and cleaning for original studies; dual screening, extraction and risk-of-bias assessment for reviews.
- 05
Analysis
The pre-specified analysis, plus sensitivity and subgroup work, certainty rating where applicable, and figure production.
- 06
Writing
Full manuscript or thesis chapters with tables, figures, supplementary files and reference list, to journal or university format.
- 07
Submission
Cover letter and portal formatting, then point-by-point response when the reviews come back — or viva preparation, if the destination is an examiner.
Authorship and transparency
What I will and will not do
Research integrity is the reason to engage a methodologist rather than a writing service. These terms are part of the engagement, not a disclaimer at the bottom of it.
-
Contributions are disclosed. Where my involvement meets the ICMJE authorship criteria I expect to be named as an author; where it does not, I am named in the acknowledgements. Undisclosed ghost authorship is not something I offer.
-
The data stays yours, and stays as collected. I analyse what you gathered. I do not generate observations, impute without stating it, or adjust data towards a preferred result, and every exclusion is recorded in the manuscript.
-
Institutional rules govern academic work. On a thesis or dissertation, your university's regulations decide what assistance is permissible. I work inside them, agree the scope in writing, and say so when a request falls outside.
-
No outcome is promised. I do not guarantee acceptance or a passed viva, and I have no arrangement with any journal, editor or indexing service. Anyone who offers you that is selling something else.
-
The work is reproducible. Search strings, master charts, extraction sheets and analysis files hand over with the project, so you or your department can re-run and defend every number in it.
-
Confidentiality runs both ways. Unpublished data, candidate and institutional identities, and manuscripts in progress are never discussed or displayed without written approval.
Engagement
Ways to work together
Every engagement starts with a written scope of work: deliverables, milestones, timeline and what sits outside the brief. Fees are quoted against that scope after a scoping call, not before it.
- Fixed scope
Full project
Topic to submission, priced on design complexity and expected sample or study count, paid against milestones.
- Partial
Analysis and figures
You have the data or the extracted set; I run the analysis, build the figures and write the results.
- Academic
Thesis support
Chapter by chapter or end to end, to your university's format and your guide's requirements.
- Ongoing
Retainer
For research cells, departments and publication teams with recurring volume — an agreed block of hours or deliverables each month.
- Remedial
Review response
A stalled manuscript, a desk rejection, or reviewer comments that need answering with new analysis.
Contact
Send me the question
Enquiries are answered within one working day. If the project is not a good fit — wrong design for the question, an unrealistic timeline, or something the literature has already settled — I will say so at the scoping stage rather than after invoicing.
drsomeshojha@gmail.comBagdogra, West Bengal, India · Remote engagements, all time zones
Useful to include
- ◆The research question, or a working title
- ◆Study design, if one has been decided
- ◆Where the project currently stands
- ◆Whether data, an extraction sheet or ethics clearance exists
- ◆Target journal or university format
- ◆Your deadline, and whether it is a hard institutional one