Most field health programs generate mountains of routine data — registers, patient cards, monthly reports — that never travel further than a district office filing cabinet. Operational research exists to close that gap: to take the data a program is already collecting and turn it into evidence that can change how the program itself is run. One of the clearest examples of this tradition in practice is SORT IT, the Structured Operational Research and Training Initiative, a model built specifically to walk field staff from a raw dataset to a published paper.
What Operational Research Actually Asks
Operational research, in the tradition this journal covers, is narrower than academic health research. It does not start with a hypothesis chosen for its scientific novelty; it starts with a question a program manager actually needs answered — why are patients being lost to follow-up, why is a vaccination round under-covering a district, why is a supply chain running out of a particular diagnostic. The method is often unglamorous: descriptive analysis of routine program records, sometimes a before-and-after comparison, rarely a randomized design. What makes it operational research rather than ordinary program monitoring is the intent to publish the finding so that other programs facing the same problem can learn from it.
This is not a niche pursuit inside global health. Frameworks for building operational and implementation research into disease-control programs have been developed jointly by bodies including the Global Fund to Fight AIDS, Tuberculosis and Malaria and TDR at WHO, precisely because donors and ministries of health kept discovering the same problem: programs that were faithfully collecting data had no reliable path for turning it into a decision. Training models exist to build that path deliberately, rather than leaving it to whichever staff member happens to have both the time and the statistical background to attempt it alone.
The SORT IT Model: Three Modules, Four Milestones
SORT IT is coordinated by TDR, the Special Programme for Research and Training in Tropical Diseases hosted at the World Health Organization, and implemented with partners including Médecins Sans Frontières and The International Union Against Tuberculosis and Lung Disease. Its structure is deliberately output-based: participants are not trained on operational research in the abstract, they are mentored while producing their own study, drawn from their own program’s data, on a fixed timeline.
The classic format runs three one-week residential modules spread across nine to twelve months, with four required milestones in between:
- Protocol. A complete operational research protocol is due by the end of Module 1.
- Instruments. Data collection tools and dummy (mock) result tables are due after Module 2.
- Dataset. A complete, cleaned dataset is due before Module 3 begins.
- Submission. Proof that the finished manuscript has been submitted to a peer-reviewed journal is due within four weeks of Module 3 ending.
Each milestone forces a decision that a less structured research process can defer indefinitely — you cannot show up to Module 2 without a protocol, and you cannot leave Module 3 without a dataset. That sequencing, more than any single statistical technique, is what the model is actually teaching.
Classic Modules vs. the Virtual Model
During the COVID-19 pandemic, the residential format that SORT IT depended on became impossible to run as designed. The response was e-SORT IT, a virtual adaptation of the same milestone structure.
| Feature | Classic SORT IT | e-SORT IT |
|---|---|---|
| Delivery | In-person residential modules | Virtual sessions |
| Structure | Three one-week modules | Same milestone sequence, remote delivery |
| Timeline | 9–12 months | Comparable timeline, pandemic-adapted |
| Mentorship | Face-to-face during modules | Remote mentoring between milestones |
What the virtual adaptation preserved is the part that seems to matter most: the fixed milestones and direct mentorship pairing each participant’s real program data with a named mentor accountable for getting it to submission.
Why the Completion Rate Is the Real Result
The headline statistic associated with the model is not a citation count but a completion rate: roughly 90 percent of participants who enter SORT IT finish by submitting their study to a peer-reviewed journal. For a research training model, that figure matters more than almost any other, because the standard failure mode of operational research capacity-building is not bad analysis — it is unfinished analysis. Field staff start a study during a slow month, and it dies the moment a new outbreak, funding cycle, or staffing gap pulls their attention elsewhere. A milestone structure with a residential or virtual deadline attached is, in effect, a mechanism against that specific failure.
What This Means for a Field Program
A program does not need to enroll in a formal course to borrow the underlying discipline. The transferable part of the SORT IT approach is the sequencing itself:
- Write the protocol and the exact indicator you will measure before touching the data — not after.
- Build the data collection tool and a mock results table before the real numbers exist, so the analysis plan cannot quietly bend to fit whatever the numbers turn out to be.
- Set a hard date for a completed, cleaned dataset, separate from the date the write-up is due.
- Commit to a submission deadline publicly, to a mentor, a supervisor, or a co-author, so the study has somewhere to go besides a shared drive.
None of this requires new equipment or a research budget most field programs don’t have. It requires treating the data a program is already recording as an asset worth finishing the analysis on — which is, at its core, what operational research in this tradition has always asked programs to do.