WF-18 — Publication and Scholarly Output

This workflow does not report publications. The Ministry already has them. KPI 4.1 is computed by MoHESR from Scopus, and KPI 4.2 is computed from SciVal. MBRU supplies the faculty details that form the denominator, and may supplement Scopus in only two narrow cases. Everything that decides IoL's score on both KPIs therefore happens before a paper is published: whether the author's ORCID record is complete, whether the affiliation string on the manuscript resolves to MBRU, and whether the author is linked to a Program Code. A paper that IoL wrote, that Scopus cannot attribute to MBRU, is invisible, and invisible in a way that no internal report will reveal, because the internal count will look correct.

1. Purpose

To ensure that every scholarly output produced by IoL faculty, staff and research students is attributed correctly to MBRU and to the programmes those authors teach on, is deposited and retained in a form that can be evidenced on demand, and rests on an authorship agreement made before the work was written rather than negotiated after it, so that the institution receives the credit for research it has actually supported and the authors receive the credit they have actually earned.

2. Scope

Scope statement. This process manages scholarly output from the authorship agreement at project inception, through affiliation and identifier hygiene, submission, acceptance and repository deposit, to the annual affiliation audit and the reconciliation of the publication and faculty records that feed KPI 4.1.

Applies to. All peer-reviewed and non-peer-reviewed scholarly output authored by IoL academic and research staff, by IoL staff holding joint appointments across Dubai Health entities, and by PGDipHPE and MScHPE learners publishing from dissertation, capstone or supervised project work. All output types carried by the HEDB Publication Code vocabulary. Both the institutional count and the programme-level attribution to PGDipHPE and MScHPE.

Does not apply to. The computation of KPI 4.1 or 4.2 themselves, which are Ministry-side. Internal reports, committee papers, teaching materials and conference attendance that produced no citable output. Grant applications, which are WF-17. Ethics approval and project registration, which are WF-15. IP disclosure, which is WF-20; note that an article in a peer-reviewed journal is expressly excluded from the KPI 4.6 copyright category, so a publication is never also an IP asset for OBEF purposes.

Applicable requirements. OBEF guide v11.5 KPI 4.1 and KPI 4.2; HEDB Data Dictionary 2026 (Publication Code, Researcher Code); CHEDS Master API specification 18 February 2026; MBRU research and authorship policy [IoL to confirm the controlling policy and its version]; publisher copyright and self-archiving terms; data protection for author personal data and identifiers.

The two definitions that govern everything below.

First, what MBRU may add to Scopus. Only two categories may be supplied outside the Scopus feed: (a) publications in Arabic, and (b) scholarly activities where traditional publication is not the primary dissemination medium, for example creative outputs in the arts. Everything else must come through Scopus. A Reasons for NON-SCOPUS entries value populated with anything outside those two grounds will be rejected. This is not a discretionary field and it is not a route for output that Scopus happens not to index.

Second, how a publication is counted. A paper with several authors from the same university counts once. A paper with authors from different universities counts once for each institution. At programme level, a faculty member working across several programmes may have the publication counted for all of them. Only outputs explicitly linked to UAE-established universities and local branches of international universities count at all.

3. Trigger, boundary and endpoint

4. SIPOC

5. Accountability

Process owner. IoL Research Lead, or the role able to set authorship expectations, require affiliation discipline and commission the annual audit. This role must be able to instruct on affiliation practice, which is a research-conduct matter, not a reporting matter.

Escalation.

6. Process steps

1. Agree authorship before the work is written. At project registration in WF-15, record the intended author group, the order convention, the criteria being applied (ICMJE or the discipline equivalent [IoL to confirm the standard MBRU adopts]), and the affiliation each author will assert. An authorship agreement made at inception prevents the dispute that a completed manuscript reliably produces, and it is the point at which the affiliation question is cheap to settle. [CAPTURE] author group, order convention, affiliation per author, agreement date. Serves 4.1 attribution and the integrity record.

2. Record an ORCID for every research-active staff member and link it to MBRU. The identifier alone is not sufficient. The ORCID record must assert the MBRU employment affiliation, because that is what allows an indexing service to resolve an ambiguous author name to this institution. [CAPTURE] Emp_ORC_ID in Employee - Basic Details.xlsx, and the ORCID employment entry. Serves 4.1 and 4.2. This is the single highest-leverage control in Pillar 4. Without it, attribution to MBRU is unreliable and the numerator is understated in a way MBRU cannot see from its own records.

3. Apply the canonical MBRU affiliation string at submission. One controlled form, published to all authors, used on every manuscript, with IoL named as the unit within it in the form MBRU has agreed [IoL to confirm the canonical string and whether an institute-level qualifier is permitted]. [CAPTURE] the affiliation string as submitted. Serves 4.1 and 4.2. Decision point. An author holding a joint appointment with a Dubai Health entity should assert both affiliations rather than choosing between them. A paper with authors from different universities counts once for each institution, so a dual affiliation does not dilute MBRU's count. Asserting only the clinical employer, however, removes the paper from MBRU's numerator entirely.

4. Record acceptance and the identifiers at the moment of acceptance. DOI, journal or publisher, year of publication, output type against the HEDB Publication Code vocabulary, number of authors, referee status, and the Scopus EID once it appears. [CAPTURE] Title of Published Work, Year of Publication, Type of Publication(Publication Code), Number of Authors, Referee Indicator, DOI, Scopus EID. Serves 4.1. Exception route. An accepted output with no EID after 90 days is flagged for the audit in step 10, not quietly moved to the non-Scopus supplement. Absence from Scopus is not one of the two permitted grounds.

5. Deposit in the institutional repository. Deposit the accepted manuscript version permitted by the publisher, with its licence, embargo end date and the deposit date recorded. [CAPTURE] deposit record, version deposited, licence, embargo. Serves the evidence base, not a KPI numerator directly. Repository deposit is genuine workflow content and should be justified on access and preservation grounds. It also happens to be the only place IoL will reliably hold the full text of its own output when an evidence request arrives.

6. Determine whether the output is a permitted non-Scopus supplement. Two grounds only: the publication is in Arabic, or it is a scholarly activity where traditional publication is not the primary dissemination medium, for example a creative output. Any other output goes through Scopus or does not count. [CAPTURE] Reasons for NON-SCOPUS entries, populated only with one of the two grounds, and left null for everything else. Serves 4.1. Decision point. A book chapter with no Scopus record, a report, a conference abstract that Scopus does not index, a preprint: none of these is a permitted ground. Record them in the IoL register for internal purposes and leave them out of the supplement.

7. Attribute the output to Program Codes. For each MBRU author, record the programmes they teach on in the reporting year. At programme level, a faculty member working across several programmes may have the publication counted for all of them, so a single paper by a member of staff who teaches on both PGDipHPE and MScHPE contributes to both programme numerators. [CAPTURE] Load_Program_codes per author in Employee - Workload.xlsx. Serves 4.1 at programme level. This step is where IoL loses programme-level publication credit today, on the assumption that Load_Program_codes is sparsely populated. That assumption should be tested before anything else in this workflow is built [IoL to confirm current population rates].

8. Confirm the denominator population and apply the exclusions correctly. Include academics, research staff, the director, deputy directors, deans and deputy deans where they carry a teaching or research load. Exclude staff below doctoral level who teach exclusively in foundation-year and general education programmes, and staff with no research duties such as technical trainers and laboratory assistants. [CAPTURE] Load_Employee_Category, Load_Full_Part_Time, Load_Research_Workload, Load_Teaching_Workload. Serves 4.1 denominator. The exclusion logic must be coded from these fields and documented, not curated by hand each year. A hand-curated exclusion list cannot be reproduced by a second analyst and cannot be defended when the ratio moves.

9. De-duplicate internal multi-author papers. A paper with several authors from the same university counts once at institution level. The register must therefore hold one institutional row per output regardless of how many MBRU authors appear on it, while retaining the per-author links needed for programme attribution. [CAPTURE] the de-duplication key (DOI, or Scopus EID where present). Serves 4.1.

10. Run the annual affiliation and misattribution audit. Search Scopus for all known MBRU name variants (including transliterations, the abbreviation, the former name if any, and misspellings), for every IoL author by ORCID and by name, and for outputs attributed to a Dubai Health entity where an MBRU affiliation was also asserted. Produce a list of missing, misattributed and duplicated records, and raise corrections with Scopus and with publishers. [CAPTURE] audit date, variants searched, records found missing or misattributed, corrections raised, corrections confirmed. Serves 4.1 and 4.2. This audit is the only lever MBRU has on KPI 4.2 and the highest-leverage control on KPI 4.1. It is a one-month task performed once a year, and it should be scheduled before the annual submission rather than after it, so that corrections have time to propagate.

11. Monitor FWCI independently. Record MBRU's and, where the tool permits, IoL's FWCI from SciVal each year, so that a discrepancy with the Ministry's figure can be raised with evidence rather than suspicion. [CAPTURE] FWCI value, source, extraction date, publication set used. Serves 4.2 monitoring only. MBRU submits nothing for 4.2.

12. Reconcile and hand to WF-26. Reconcile the IoL register against Institute - Publications.xlsx, Institute - Scholarly Output.xlsx and the faculty denominator, confirm that every research-active staff member has an ORCID, and release the extract with the exclusion logic documented. [CAPTURE] reconciliation date, unresolved differences, exclusion logic version. Serves 4.1.

7. OBEF data generated

Capture rule. Three things cannot be reconstructed afterwards and must be recorded at the time: the affiliation string as it was submitted (once a paper is published, the affiliation is fixed and only a publisher correction changes it), the authorship agreement and its date (reconstructed afterwards it is a negotiation, not a record), and the programme attribution for the reporting year in which the output appeared (staff move between programmes, and a later workload record will not describe the year the paper was published). Everything else in the table above can be recovered from the DOI.

Reproducibility test. A second analyst cannot reproduce KPI 4.1 from IoL's records alone, and this is a structural feature rather than a defect: the numerator lives in Scopus and the Ministry computes it. What the analyst must be able to reproduce is the denominator (from Employee - Workload.xlsx with documented exclusion logic), the non-Scopus supplement (from Institute - Publications.xlsx filtered to the two permitted grounds), and the programme attribution (from Load_Program_codes). If any of those three requires asking someone what happened, the capture defect is real even though the KPI is Track B. On current assumptions, Load_Program_codes is the one that fails.

8. Max-score design

What this workflow must do to reach the target.

1. Get every research-active IoL staff member an ORCID with an MBRU employment record on it. Not a list of identifiers in a spreadsheet: an ORCID record that asserts the affiliation. This is the only lever on 4.2 and the largest single lever on 4.1.
2. Publish one canonical affiliation string and require it. Variation in the affiliation string is the mechanism by which a small institution's output disperses across several apparent institutions in an indexing service.
3. Audit Scopus annually for MBRU name variants and misattributions, and raise corrections. Schedule the audit before the submission, not after.
4. Populate Load_Program_codes for every research-active staff member. Without it, programme-level 4.1 for PGDipHPE and MScHPE is understated or null, and the 3.5% programme weight is at risk while the 2.5% institutional weight is unaffected. The programme level is where the larger weight sits.
5. Code the denominator exclusions from Load_Employee_Category and Load_Research_Workload and document the logic, so the same rule is applied every year and can be shown to a reviewer.
6. Populate Reasons for NON-SCOPUS entries only on the two permitted grounds, and leave it null otherwise. A rejected supplement costs credibility on a submission where credibility is the asset.
7. Instruct dually appointed staff to assert both affiliations. A paper with authors from different institutions counts once for each, so dual affiliation is additive, not divisive.

[REDESIGN] actions.

Sequencing note. R1 and R2 must precede R5, because an audit run before the identifiers and the affiliation standard exist will find problems it cannot fix and will have to be run again. R3 is independent of the other five and is the one with a programme-level weight attached, which makes it the first thing to do if only one action is taken.

A caution about the size of the prize. The score maximisation plan models action A12, the ORCID and affiliation audit, at +0.3 points. That is a correct estimate of the modelled gain and a poor description of the risk the action removes. The modelled gain is small because the model assumes attribution is broadly working. If it is not working, the loss is a fraction of the whole of 4.1 and 4.2, and nobody at IoL would be able to see it, because the internal publication list would still be complete. R1, R2 and R5 are insurance against an invisible loss, and should be argued that way rather than as a score improvement.

9. Indirect strategy where data cannot be collected

Category 1 applies to KPI 4.2 in full and to KPI 4.1 in part. Category 4 applies to programme attribution. Category 2 is available for 4.1 and should be understood before it is used.

Category 1, the Ministry holds the instrument. KPI 4.2 is computed by MoHESR from SciVal, and MBRU submits nothing at all. KPI 4.1 is computed from Scopus, and MBRU submits only the faculty denominator and the two categories of permitted supplement. In both cases the lever is the input, not the number. The input is affiliation and identifier hygiene, and it is the same input for both KPIs, which is the single most useful fact in this workflow: one control, ORCID plus a canonical affiliation plus an annual audit, serves 4.0 percentage points of the framework at institutional level and 5.0 at programme level.

There is no second lever on 4.2. Do not spend pipeline effort on it beyond the hygiene that 4.1 requires anyway, and do not present FWCI as a managed number in any internal report. It is monitored, not managed.

A note on the 4.2 band structure, because it changes what a movement is worth. The anchors are 0 / 0.8 / 0.9 / 2 / 4. A publication set at world average (1.0) sits just inside the Medium band. Moving from 0.85 to 0.95 crosses a band boundary and is worth roughly 25 points of the KPI; moving from 1.5 to 1.6 is worth roughly 2. The band structure means the score is most sensitive exactly where most institutions sit, so a small number of poorly attributed papers near world average can move the score more than a genuine improvement in citation performance further up the range. That is an argument for attribution accuracy, not for gaming: the point is that the number is noisy at IoL's likely operating range and should not be treated as a management signal year to year.

Category 4, the activity exists but is invisible. This is the programme-level case. IoL faculty publish. Whether those publications reach the PGDipHPE and MScHPE programme rows depends entirely on whether Load_Program_codes links the author to the programme for the year in question. Nothing about the research has to change. The response is capture, at the start of the academic year, in the workload record, and it is action R3.

The same category covers the dually appointed author. IoL sits inside Dubai Health, and staff who hold joint clinical and academic appointments will often assert only the clinical affiliation on a manuscript, because that is the identity under which the clinical work was done. The output exists, it is MBRU's by the OBEF counting rules, and it is invisible because of a habit rather than a rule. The response is R2 and an instruction, not a tactic.

Category 2, substitute data, is available and needs care. The OBEF guide explicitly invites institutions to submit their own data directly to MoHESR for KPI 4.1 where they consider it more accurate than the standard route. That invitation is real and IoL should know it exists. It is also not the same as the two permitted non-Scopus grounds, and the two must not be confused: the supplement fields in Institute - Publications.xlsx accept only Arabic-language and non-traditional-dissemination output, whatever a substitute-data submission might separately contain.

Before relying on Category 2 for 4.1, resolve the ambiguity rather than assuming a reading. The guide states in one place that MBRU may supplement Scopus in only two cases, and in another that institutions may submit their own data for 4.1 where it is more accurate. Those two statements are not obviously compatible. The conservative reading, which this workflow adopts until MoHESR says otherwise, is that the two permitted grounds govern what enters the publication templates, and that any broader substitute-data submission is a separate, negotiated act requiring the Ministry's agreement in advance. Put the question in writing, record the answer, and apply it consistently.

Category 3 does not apply and must not be claimed. IoL produces scholarly output. There is no absence to classify and no redistribution available on 4.1 or 4.2 at either level.

Category 5, if the ratio is genuinely low. If the audit confirms the attribution is sound and the ratio is still at or below 2 per FTE, then the performance is what it is. Report it, and open a corrective action plan through WF-04. Do not compensate by widening the non-Scopus supplement, which is the specific temptation this KPI creates.

Boundary check. This workflow must never:

• populate Reasons for NON-SCOPUS entries with anything other than Arabic-language publication or non-traditional dissemination, and specifically never with "not indexed in Scopus", "conference abstract", "book chapter", "preprint" or "report";
• count a publication for MBRU where no author asserted an MBRU affiliation and none can be evidenced;
• count a single paper more than once at institution level because it has several MBRU authors;
• remove staff from the KPI 4.1 denominator on any basis other than the two stated exclusions, namely sub-doctoral staff teaching exclusively in foundation and general education, and staff with no research duties;
• add staff to the denominator, or exclude them, after seeing the effect on the ratio;
• claim a publication for a programme where the author does not teach on that programme in the reporting year;
• count output linked to an institution that is neither UAE-established nor a local branch of an international university;
• record a student output under a supervisor's name, or a supervisor's output under a student's, in order to move it between KPI 4.1 and KPI 4.4;
• present a claimed publication as a KPI 4.6 IP asset, since journal articles are expressly excluded from the copyright category;
• report a SciVal FWCI figure as MBRU's OBEF result, since the Ministry computes 4.2 on its own extraction and the two will not necessarily agree.

10. Service standards

11. Records and evidence

Appendix B readiness. Today, IoL could most likely produce a list of its publications and the DOIs within 15 working days. It could not produce, on current assumptions: an ORCID completeness report, an affiliation audit, a documented denominator exclusion rule, or a programme attribution per output [IoL to confirm]. Of those four, the denominator exclusion rule is the one a reviewer is most likely to ask for, because the denominator is the only part of KPI 4.1 that MBRU actually submits.

12. Risks and controls

13. Performance measures

The first four measures are the ones that decide this workflow's contribution. IoL can publish well and score poorly on 4.1, and the difference is entirely in these four rows.

14. Change control