Knowledge Brokering

Knowledge Brokering in Health Technology Assessment and Health Economics

A knowledge broker is a person or organisation positioned between the people who produce research and the clinicians, managers and policy makers who decide what a health system does. It matters to health economists because a cost-effectiveness result changes nothing until a payer, committee or clinician acts on it, and HTA bodies such as NICE have been described as knowledge brokers. Trials give a mixed picture of whether brokering raises evidence use, and value of implementation methods set an upper bound on what a brokering programme is worth.

Why research and health service decisions drift apart

Lomas, writing in 2007 as head of the Canadian Health Services Research Foundation (CHSRF), argued that research and health services were poorly connected worlds. Universities reward grants and publication more than synthesis and engagement with services, while health system governance rewards consensus more than the use of research. Researchers tend to treat a decision as a single event, and decision makers treat research as a product to be bought off the shelf.

His proposed remedy was linkage and exchange between the processes that create research findings and the processes that weigh values in decision making. He cited a systematic review by Innvaer and colleagues which concluded that personal two-way communication between researchers and decision makers should be used to promote research use. An updated review by Oliver and colleagues in 2014, covering 145 new studies, found that the barriers to evidence use by policy makers reported most often were poor access to good quality relevant research and a lack of timely research output, and the facilitators reported most often were collaboration and improved relationships and skills.

The CHSRF definition and three broker roles

CHSRF defined knowledge brokering in 2003 as all the activity that links decision makers with researchers and helps them understand each other's goals and professional cultures, influence each other's work, form new partnerships and promote the use of research-based evidence in decisions. Ward, House and Hamer noted in 2009 that brokerage can reside in individuals, organisations or structures.

Ward and colleagues grouped brokering into three models. A systematic review by Bornbaum and colleagues in 2015, covering 22 studies in health-related settings, found support for all three roles.

HTA agencies and economic evidence as a brokering problem

Fenwick, Claxton and Sculpher noted in 2008 that issuing guidance on cost-effective technologies does not by itself lead to cost-effective service provision, because implementation is rarely perfect. Among the causes they listed were different incentives facing practitioners and an asymmetry of information between policy makers and local decision makers.

Health technology assessment agencies sit at this junction. Lomas named NICE as an exception to the general disconnection between research and services. Tantivess and colleagues, writing from HITAP in Thailand and NICE International in 2017, stated that the role of independent institutions such as NICE in translating evidence into policy, as knowledge brokers, had been highlighted before.

The same institutional model has been taken to low and middle income countries. HITAP, established in 2007 as a semi-autonomous research arm of the Thai Ministry of Health, has its cost-effectiveness and budget impact analysis work formally embedded in decisions on the reimbursable medicines list and the benefit package of the national universal health coverage scheme. Tantivess and colleagues concluded from HITAP and NICE International programmes in India, Colombia, Myanmar, the Philippines and Vietnam that management, coordination and communication capacity mattered for the generation and use of HTA evidence, alongside analytical techniques such as economic evaluation.

The international Decision Support Initiative (iDSI) was established in 2013 as a collaborative platform to speed progress on evidence-informed priority setting, particularly in low and middle income countries. A ten-year review by Baker and colleagues reported that iDSI supported regional networks in Asia and Africa and developed HTA capacity. Its lessons included sustained political will, a clear mandate, context-appropriate methods such as adaptive HTA, and donor-funded programmes that integrate with national systems instead of undermining them. The review also reported improved decisions on service provision, procurement and clinical care.

Valuing brokering with the expected value of perfect implementation

Fenwick and colleagues set out a framework in which both information and implementation can be at their current level or perfect. The expected value of perfect implementation (EVPIM) is the difference between the expected value of a decision implemented perfectly and the expected value with implementation at its current level, measured in net monetary benefit. The population EVPIM is an upper bound on the return to any strategy that changes implementation, including a brokering programme, and so gives a necessary condition for such a strategy to be worthwhile.

For a technology with a positive expected incremental net monetary benefit, and assuming every extra patient treated gains the average benefit, the annual population EVPIM simplifies to the expression below.

$$EVPIM = N (1 - r_0) B$$

where $N$ is the eligible population per year, $r_0$ is the current proportion of eligible patients who receive the technology, and $B$ is the expected incremental net monetary benefit per patient at the decision maker's threshold.

A specific brokering programme raises implementation from $r_0$ to $r_1$ at a cost, and its expected net value is as follows.

$$V = N (r_1 - r_0) B - C$$

where $V$ is the net value of the programme per year, $N$, $r_0$ and $B$ are as above, $r_1$ is the proportion treated after the programme, and $C$ is the annual cost of the programme. For a technology that is not cost-effective, perfect implementation means $r = 0$, and the same logic values brokering that reduces use. In one of the Fenwick case studies, prophylactic removal of wisdom teeth, the value of further research was negligible while the value of strategies to change practice was substantial.

Worked example with a recommended technology and a broker programme

All figures here are illustrative. A recommended technology gives 0.05 extra QALYs per patient at an extra cost of GBP 400, and the decision maker's cost-effectiveness threshold is GBP 20,000 per QALY. Each year 10,000 patients are eligible, and 40% currently receive it.

$$B = 20000 \times 0.05 - 400 = 1000 - 400 = 600$$

where $B$ is the incremental net monetary benefit per patient in GBP, 20,000 is the threshold, 0.05 is the QALY gain and 400 is the incremental cost.

$$EVPIM = 10000 \times (1 - 0.40) \times 600 = 10000 \times 0.60 \times 600 = 3600000$$

where 10,000 is the eligible population, 0.40 is current uptake and 600 is $B$. At most GBP 3.6 million of net benefit a year, about 180 QALYs at the threshold, can be recovered by any implementation strategy.

A regional broker programme costs GBP 250,000 a year. The break-even rise in uptake is GBP 250,000 divided by GBP 6,000,000 (10,000 times 600), or about 4.2 percentage points.

The programme is good value only if it moves uptake by more than about four points, and a programme with no effect on uptake would displace 12.5 QALYs elsewhere in the budget. The broker's effect on practice is the decisive and least certain parameter, so it belongs in sensitivity analysis.

What trials and reviews show about effectiveness

Dobbins and colleagues ran a randomised trial in Canadian public health departments from 2004 to 2006, with interventions delivered over one year in 2005. Of 141 health departments, 108 (77%) took part, and they were randomly allocated to access to an online registry of research evidence, tailored targeted messages, or a knowledge broker. There was no significant effect on the primary outcome, the extent to which research evidence was used in a recent programme decision. For the number of evidence-informed healthy body weight policies and programmes, only tailored messages had a significant effect.

Organisational research culture moderated the result. Departments with low research culture that received the broker showed a significant improvement in policies and programmes, while those with high research culture showed no benefit and possibly a decrease. The authors suggested that brokers may build capacity where evidence use is weak.

The Bornbaum review examined outcome data from eight studies and found only two of acceptable methodological rigour, so it judged the evidence on broker effectiveness inconclusive. Ward and colleagues described a trial by Amsallem and colleagues in which brokers discussing cardiovascular evidence with clinicians changed intent to prescribe but not actual prescribing. Lomas reported that managers in a CHSRF training programme who said they used research most or all of the time rose from 21% to 50% over two years, a self-reported figure without a control group.

Knowledge brokering beside related ideas

Knowledge translation is the wider process of moving research into practice and policy, and brokering is one strategy within it, alongside tailored messages and evidence registries as in the Dobbins trial. Ward and colleagues noted that passive one-way dissemination is widely acknowledged as ineffective, whereas brokering relies on repeated interaction. Decision support refers to software or structured tools used at the point of choice, and diffusion of innovation is Rogers' theory of how new practices spread, which brokering tries to speed up.

Within HTA the distinction is between producing evidence and getting it used. An agency's economic evaluation and appraisal produce a recommendation, and the brokering work lies in agreeing the questions with decision makers, explaining the results and supporting implementation. Value of information asks whether more research is worth funding, and the value of implementation asks whether changing practice is worth funding, and Fenwick and colleagues showed the answers can differ between technologies.

Where knowledge brokering is overrated or misread

One misunderstanding is that more evidence use always means better decisions. In economic terms the yardstick is the change in net benefit, so a broker who raises uptake of a technology that is not cost-effective reduces population health.

Another is that brokering is a proven intervention. Ward and colleagues identified the lack of knowledge about how brokering works, which contextual factors influence it and how effective it is as the biggest challenge. They noted that it is time-consuming and hard to evaluate because models are mixed within one programme, and cited the suggestion that one year is probably too short for significant gains in capacity, which matters when trials last a single year.

Independence is a further concern. Ward and colleagues noted that a consultant-client relationship can call into question how impartial a broker can be, and an HTA body that both produces and promotes its own recommendations faces a similar tension. In low and middle income countries the iDSI review stressed politically sensitive and country-led support, and an outside broker that bypasses national institutions risks leaving no lasting priority setting process behind.

Sources

• Lomas J. The in-between world of knowledge brokering. BMJ 2007;334(7585):129-132. https://doi.org/10.1136/bmj.39038.593380.AE
• Ward V, House A, Hamer S. Knowledge brokering: the missing link in the evidence to action chain? Evidence and Policy 2009;5(3):267-279. https://doi.org/10.1332/174426409X463811
• Bornbaum CC, Kornas K, Peirson L, Rosella LC. Exploring the function and effectiveness of knowledge brokers as facilitators of knowledge translation in health-related settings: a systematic review and thematic analysis. Implementation Science 2015;10:162. https://doi.org/10.1186/s13012-015-0351-9
• Dobbins M, Hanna SE, Ciliska D, et al. A randomized controlled trial evaluating the impact of knowledge translation and exchange strategies. Implementation Science 2009;4:61. https://doi.org/10.1186/1748-5908-4-61
• Oliver K, Innvar S, Lorenc T, Woodman J, Thomas J. A systematic review of barriers to and facilitators of the use of evidence by policymakers. BMC Health Services Research 2014;14:2. https://doi.org/10.1186/1472-6963-14-2
• Fenwick E, Claxton K, Sculpher M. The value of implementation and the value of information: combined and uneven development. Medical Decision Making 2008;28(1):21-32. https://doi.org/10.1177/0272989X07308751
• Tantivess S, Chalkidou K, Tritasavit N, Teerawattananon Y. Health technology assessment capacity development in low- and middle-income countries: experiences from the international units of HITAP and NICE. F1000Research 2017;6:2119. https://doi.org/10.12688/f1000research.13180.1
• Baker P, Barasa E, Chalkidou K, et al. International partnerships to develop evidence-informed priority setting institutions: ten years of experience from the international Decision Support Initiative (iDSI). Health Systems and Reform 2023;9(3):2330112. https://doi.org/10.1080/23288604.2024.2330112