WF-17 — Research Funding, Industry and International Partnership

The material clause for MBRU sits in the approved partner-type list. Industry collaboration for KPI 4.3 admits healthcare institutions (a hospital, clinic or healthcare provider), research centres, professional associations and non-profit organisations, while excluding federal ministries and other public bodies, which are classed as government agencies. For an institute of learning inside an academic health system, health-system partners plausibly qualify as healthcare institutions. That is a large and entirely legitimate lever, and it depends on one thing: recording partner type at project launch.

Both KPIs are cumulative over five years, both count a collaboration once, in its launch year, and both divide by total research projects. If 4.3 and 5.4 use different denominators, one of them is wrong. That is cross-KPI assertion C2, and it is the fastest check available on whether this workflow is working.

1. Purpose

To ensure that every collaboration IoL enters into with an industry or international partner is tested against the qualifying routes at the moment it forms, classified against the approved partner-type vocabulary, evidenced by an agreement demonstrating the partner organisation's awareness, valued where money or in-kind contribution is involved, and counted exactly once in its launch year across a five-year window, so that two KPIs worth 7.0 programme points are reported from a register rather than assembled from memory at submission time.

2. Scope

Scope statement. This process manages the industry and international partnership attributes of IoL research from the point at which a collaboration is contemplated, through qualification testing and valuation, to the once-only counting of the collaboration in its launch year and the reconciliation of its financial value against every other place that value appears.

Applies to. Every registered IoL research project with an industry partner, an international partner or both, whether the contribution is financial, in-kind or intellectual; and the doctoral joint-supervision relationships that qualify for KPI 5.4 through WF-16. It applies to health-system partners inside Dubai Health as much as to external and overseas ones, and to unfunded collaborations, which qualify through the intellectual route.

Does not apply to. Registering the project itself, which is WF-15 and is a hard precondition here. It does not manage the negotiation or signature of agreements (WF-24), publication of the resulting output (WF-18), impact roll-up (WF-19), IP (WF-20), or teaching partnerships and co-delivered courses, which are KPI 3.3 and belong to WF-03. A partner who teaches is not thereby a research partner, and a partner who funds research is not thereby a co-deliverer of a course. The two claims are separate, evidenced separately and must not be transferred between each other.

Applicable requirements. OBEF guide v11.5 KPI 4.3 and 5.4, and KPI 3.4 where a contribution has a financial value; the approved Partner Type and Partnership Type vocabularies in the HEDB lookups; CHEDS/HEDB Institute - Research Projects.xlsx, Institute Partnerships.xlsx and Students - Research.xlsx; Appendix B evidence provisions for partnership documents and proof of contribution; MBRU financial regulations for the valuation of in-kind contributions.

The two KPIs side by side, because their differences are where the points are lost.

Read the anchors against each other. KPI 4.3 needs 65% of projects to reach the High band and 75% to score 100, which is demanding. KPI 5.4 needs only 40% for the High band but 80% for 100, so the top of that KPI is out of reach for most institutions and the realistic target is the High band. The third route is what makes 5.4 tractable at all, and it exists only for 5.4.

The awareness requirement, common to both. The partner institution must be aware of the collaboration. Working with an individual affiliated to an organisation, without that organisation's demonstrable knowledge or approval, does not qualify. This is not a formality: it is the difference between a partnership and a collegial relationship, and the evidence for it is an agreement, an MoU or a documented institutional approval that existed at the time.

3. Trigger, boundary and endpoint

4. SIPOC

5. Accountability

Process owner. IoL Director of Research, or the role able to determine partner classification and to require evidence before a claim is made.

Escalation.

6. Process steps

1. Confirm the project is registered. No Project ID, no launch year, no claim. Where a collaboration exists on unregistered work, route it to WF-15 and register the project first. [CAPTURE] Project ID and launch year, inherited, not created here. Serves 4.3, 5.4 and assertion C6. This workflow never invents a project. It classifies projects that already exist in the register. That separation is what makes the denominator trustworthy: the set of projects is fixed by WF-15 before anybody knows which of them will qualify.

2. Classify the partner type against the approved vocabulary. Closed list, no free text. [CAPTURE] Partner Type for an industry partner, International Partner Type for an international one, and the partner's identity. Serves 4.3 and 5.4. The approved industry partner types include healthcare institutions, meaning a hospital, clinic or healthcare provider, research centres, professional associations and non-profit organisations. Federal ministries and other public bodies are excluded; they are government agencies and do not qualify as industry partners. Decision point, and it is the most consequential classification in this workflow. Dubai Health entities that are hospitals, clinics or healthcare providers plausibly meet the healthcare-institution definition on a plain reading. Entities that are ministries, regulators or public administrative bodies do not. Classify each entity once, on its actual character, with a written rationale and a second signature from Quality and IQA, and hold the classification constant across projects and years. A classification that varies by project is not a classification, it is an outcome. [IoL to confirm] the legal and organisational status of each Dubai Health entity IoL partners with, because the answer is a matter of fact about those entities and not a matter of judgement about the KPI.

3. Test the financial route. Both conditions, not either: the partner contributes at least 20% of the project budget, and the project budget is at least AED 50,000. Valued in-kind contributions count towards the 20% if both thresholds are met. [CAPTURE] Funding Amount, Partner_funding or International_Partner_funding, Type of industry contribution or Type of international partner contribution, and the computed ratio. Serves 4.3 and 5.4 route 1, and feeds 3.4. Exception route. A project with a budget below AED 50,000 fails this route regardless of the partner's share, even at 100%. Do not attempt to aggregate several small projects to clear the threshold; the test is per project.

4. Test the intellectual route. Two conditions, both required: the project is linked to a tangible output, being a publication, a product or IP; and the partner's contribution is verifiable, for example through a Scopus-indexed co-publication or jointly owned IP. [CAPTURE] Intellectual Contribution, Intellectual Contribution_other, International_Partner_Intellectual Contribution, and the identifier of the linked output. Serves 4.3 and 5.4 route 2. Decision point. A claimed intellectual contribution with no linked tangible output is not submittable. This is the most tempting shortcut in the workflow, because collegial intellectual collaboration is real and constant while the tangible output may be years away. The rule is not that the collaboration is not real; it is that the KPI counts the evidenced version of it. Where the output is pending, hold the claim until the output exists and then count the project in its launch year, not in the year the output appeared.

5. Test the doctoral joint-supervision route, for KPI 5.4 only. Where a doctoral student is jointly supervised with a named co-supervisor at an international partner organisation, the collaboration qualifies without any financial or intellectual test. [CAPTURE] from Students - Research.xlsx: Name of the international partner organisation, Name of the co-supervisor from the international partner organisation, Student Level(Level Code), Type of research. Serves 5.4 route 3. This route does not exist for KPI 4.3 and is easy to overlook. It is also the cheapest numerator available anywhere in Pillar 5, and the one most commonly forgotten, because the relationship lives in a student record rather than a project record. WF-16 hands these over; this step is where they are claimed.

6. Confirm partner awareness evidence. For every claimed collaboration, an agreement, MoU or documented institutional approval that evidences the partner organisation's knowledge of the collaboration. [CAPTURE] Agreement/MOU Number on the project record, cross-referenced to the agreement held by WF-24. Serves 4.3 and 5.4 counting rules. Note the template placement, which is easy to get wrong: Agreement/MOU Number is a column of Institute - Research Projects.xlsx, not of Institute Partnerships.xlsx. The partnerships template carries Agreement type instead.

7. Value in-kind contributions. Where the financial route depends on in-kind contribution, the valuation must be documented: an audit report, an independent third-party valuation or a formal market valuation. [CAPTURE] valuation amount, method, source and date. Unvalued in-kind is excluded, not estimated. An estimate produced internally at submission time is not a valuation, and Appendix B specifically names third-party and market valuations as the evidence for this.

8. Set the indicators. [CAPTURE] Joint Industry Research Indicator, Industrial partner name, International Research Collaboration Indicator, International partner name, and the qualifying route recorded against each. Serves 4.3 and 5.4 numerators. Record the route, not just the flag. A Y with no route is unauditable, and the route is the first thing an evidence request will ask for.

9. Record the collaboration once, against the project's launch year. Write to the Launch-Year Register maintained by WF-15. [CAPTURE] the launch-year entry with its route and evidence references. Serves assertion C6. Decision point, and it recurs every year. A five-year cumulative window means a project counted in 2024 stays in the window until it rolls out, and the temptation at each submission is to re-count it as current. A project is never re-counted in a later year of the window. The register is the control, and it only works if it is consulted before the extract rather than after.

10. Confirm the denominator is identical for both KPIs. The denominator for 4.3 and for 5.4 is all Frascati-qualifying research projects over the same five-year window, taken from WF-15. [CAPTURE] the denominator with its construction rule and extract date. Serves assertion C2. If the two numbers differ, one of them is wrong, and the difference is almost always that one numerator's population was used as the other's denominator, or that a different window was applied. Compute the denominator once and use it twice.

11. Reconcile the contribution value across templates, for KPI 3.4. A single industry research grant can appear in Institute - Financials.xlsx as revenue, in Institute - Research Projects.xlsx as Partner_funding, and in Institute Partnerships.xlsx as Value of PARTNER Contribution towards Partnership. [CAPTURE] the authoritative source declared for this contribution type, and the reconciliation result. Serves 3.4 and assertion C9. Define one authoritative source per contribution type and reconcile the others to it. Research grant income is naturally authoritative in the project record; institutional revenue lines and partnership values are then derived views of it, not independent counts. Declare which is which once, in writing, and hold it.

12. Release to submission with the numerators, the shared denominator, the route record and the evidence references. [CAPTURE] the extract, its date, and the assertion results for C2, C6 and C9.

7. OBEF data generated

Capture rule. Five things cannot be reconstructed once the moment passes. The partner type, because an organisation recorded as free text three years ago cannot be classified against a closed vocabulary without a fresh judgement that nobody can evidence. The partner's funding share against the project budget, because both numbers change over a project's life and only the values at the point of agreement test the 20% and AED 50,000 thresholds. The in-kind valuation with its method and source, because a valuation produced later is an estimate. The agreement or MoU reference evidencing awareness, because either a document existed at the time or it did not. The launch year, inherited from WF-15, because the entire once-only counting rule rests on it.

Reproducibility test. With the launch-year register and the partner attributes in place, a second analyst reproduces both KPIs by filtering the project register on the five-year window, counting distinct projects for the denominator, and counting those with a qualifying route and evidence for each numerator. The route field is what makes the numerator repeatable rather than a judgement re-made each year. Today, neither KPI is reproducible at IoL, because partner type, contribution share and launch year are not held as fields. The specific failure is not that IoL has no industry partners. It is that IoL cannot demonstrate which projects had them and when.

8. Max-score design

Read the two anchor sets against each other before allocating effort. KPI 4.3 is worth 5.0 programme points and its High band opens at 65%, which is a demanding ratio requiring most of the portfolio to be partnered. KPI 5.4 is worth 2.0 points and its High band opens at 40%. Effort spent bringing 4.3 from 45% to 65% is worth roughly 1.0 point; effort spent bringing 5.4 from 25% to 40% is worth roughly 0.4. Both are worth having, and the first is worth having first.

What this workflow must do to reach the target.

1. Classify every partner organisation once, against the closed vocabulary, with a written rationale. For Dubai Health entities this is the single highest-value classification decision in the workflow and it must be made on the entities' actual character, not on what would help.
2. Apply the intellectual route deliberately, because most educational research is unfunded and the financial route's AED 50,000 floor will exclude much of it. The intellectual route has no budget threshold. What it has is a tangible-output requirement, which means this KPI depends on WF-18 delivering publications and on those publications naming the partner.
3. Claim the doctoral joint-supervision route for 5.4, sourced from WF-16, every year. It is free numerator that already exists somewhere in the student records.
4. Compute the denominator once and use it for both KPIs, then assert C2.
5. Consult the launch-year register before every extract, then assert C6. A five-year window plus once-only counting fails silently in both directions without it.
6. Declare one authoritative source per contribution type and reconcile the other two, then assert C9.
7. Record the route and the evidence reference against every Y indicator, so a claim can be defended rather than re-argued.

[REDESIGN] actions.

Sequencing note. R1 gates the value of everything else, because if health-system partners are not classifiable as healthcare institutions then 4.3's realistic target falls a long way and effort should move to the intellectual route and to 5.4. Establish the classification first, on the facts, and let it determine where the remaining effort goes. R6 has the decaying window: four of the five cumulative years are already in the past.

9. Indirect strategy where data cannot be collected

Category 4, the activity exists but is invisible, is the dominant case, and Category 3 applies at the margin.

Category 4. IoL conducts educational research with health-system colleagues, professional bodies and overseas collaborators. Some of it is funded, most of it is not, much of it produces publications, and almost none of it is recorded as a partnered project with a partner type, a contribution route and an agreement reference. The response to Category 4 is capture, not tactics, and here the capture is specific: partner type at launch, route at qualification, evidence reference at claim. Nothing in this paragraph requires IoL to do more research or to find new partners. It requires the partners IoL already has to be recorded as partners.

The legitimate lever, stated as precisely as it deserves. The approved partner-type list for industry collaboration includes healthcare institutions, research centres, professional associations and non-profit organisations, and excludes federal ministries and other public bodies. For an institute of learning inside an academic health system, a substantial proportion of research collaborators are hospitals, clinics and healthcare providers, which is exactly what the healthcare-institution type describes. Recording them as such is classification, not manoeuvre, provided three conditions hold: the entity really is a healthcare provider rather than a public administrative body; the classification is made once and applied consistently regardless of whether it helps in a given case; and the collaboration itself qualifies through one of the routes with its evidence. Fail any of the three and the claim moves to the wrong side of the line drawn in the charter.

Category 3, the activity does not exist, applies narrowly. Where IoL genuinely has no international partner on any project in the window, KPI 5.4 reports zero and no redistribution is available for it, since redistribution exists at programme level only for 3.4, 5.2, 5.3 and 6.2. A zero on 5.4 is reported as a zero. The response is not a classification argument, it is R3 and R5: find the doctoral co-supervision that already exists, and formalise the relationships that already run.

A structural point about the denominator, carried over from WF-15 and repeated because it bites here. Both KPIs are ratios over total research projects, so the ratio improves if the denominator shrinks. This workflow does not control the denominator, WF-15 does, and that separation is deliberate. The correct response to a low ratio is to evidence more collaborations, not to register fewer projects. Assertion C2 compares the two denominators, and an Appendix B request for the project list compares both against reality.

Ambiguities in the guide, stated rather than resolved.

1. Whether a Dubai Health entity is a healthcare institution or a government agency. The vocabulary distinguishes healthcare institutions (hospital, clinic, healthcare provider) from government agencies (federal ministries and other public bodies). A government-owned hospital sits between the two readings on the face of the text. This is worth a written question to MoHESR, because the answer materially changes KPI 4.3 for every clinical university in the UAE, and because a classification confirmed in writing is worth far more than the same classification asserted. Until an answer arrives, classify on the entity's function, record the rationale, and disclose the basis in the Mechanism of collecting the data field.
2. How a partner joining after the launch year is treated. The project is counted once in its launch year, so a partner arriving in year three either qualifies the project retrospectively in its launch year or does not qualify it. The guide does not say. This workflow's working assumption, stated as an assumption: the collaboration is counted in the project's launch year, because the counting unit is the project, and the date the partner joined is recorded so the treatment is visible. Confirm it.
3. Whether "total research projects" means projects launched in the window or projects active in it. A cumulative five-year window with once-only launch-year counting implies launched-in-window for the numerator. If the denominator were active-in-window, the two would be inconsistent and the ratio would be understated. Apply launched-in-window to both, record the choice, and hold it constant across years.
4. Whether an in-kind contribution of staff time from a health-system partner can be valued at all. MBRU financial regulations may not admit a valuation method for seconded clinical time. [IoL to confirm] with Finance. Where no defensible valuation exists, the contribution is excluded from the financial route and the collaboration is tested against the intellectual route instead, which is often the better fit.

Boundary check. This workflow must never:

• record a partner as aware of a collaboration without an agreement, MoU or documented institutional approval to evidence it;
• claim an intellectual contribution with no linked tangible output, however genuine the intellectual collaboration;
• re-count a project in a later year of the five-year window, for either KPI;
• classify an organisation as a healthcare institution, research centre, professional association or non-profit when it is in fact a federal ministry or other public body;
• vary a partner's classification between projects or between years according to which classification helps;
• treat a co-authorship as a project, or a teaching partnership under KPI 3.3 as a research partnership under 4.3;
• aggregate project budgets to clear the AED 50,000 minimum, which is a per-project test;
• estimate an in-kind valuation internally in place of an audit report, third-party valuation or formal market valuation;
• use different denominators for 4.3 and 5.4, or use a numerator population as a denominator;
• count the same contribution value in Institute - Financials.xlsx, Institute - Research Projects.xlsx and Institute Partnerships.xlsx as three contributions;
• claim a Master's-level international co-supervision under the doctoral route without a Ministry answer permitting it.

10. Service standards

11. Records and evidence

Appendix B readiness. The Ministry can demand partnership documents, proof of financial contribution in specified forms, proof of intellectual contribution, and partner contact data, without notice. Today IoL could produce some agreements and few valuations within 15 working days, and could not produce a qualification test record for any project, because none exists. The gap is not the documents, which largely exist somewhere; it is the index that connects a claim to the document that supports it. That index is the qualification test record at step 4, and it is created in the same minutes as the claim itself.

12. Risks and controls

13. Performance measures

The three integrity measures are the ones that will be tested by somebody else. C2 in particular is checkable by anybody with both submissions in front of them, takes one subtraction, and a failure is not a rounding difference: it is evidence that the two KPIs were assembled separately, which invites a look at everything else.

14. Change control