Health Service Delivery and Economic Evaluation of Limb Lower Bone-Anchored Prostheses: A Summary of the Queensland Artificial Limb Service's Experience

The emergence of skeletal prosthetic attachments leaves governmental organizations facing the challenge of implementing equitable policies that support the provision of bone-anchored prostheses (BAPs). In 2013, the Queensland Artificial Limb Service (QALS) started a five-year research project focusing on health service delivery and economic evaluation of BAPs. This paper reflects on the QALS experience, particularly the lessons learned. QALS' jurisdiction and drivers are presented first, followed by the impact of outcomes, barriers, and facilitators, as well as future developments of this work. The 21 publications produced during this project (e.g., reimbursement policy, role of prosthetists, continuous improvement procedure, quality of life, preliminary cost-utilities) were summarized. Literature on past, current, and upcoming developments of BAP was reviewed to discuss the practical implications of this work. A primary outcome of this project was a policy developed by QALS supporting up to 22 h of labor for the provision of BAP care. The indicative incremental cost-utility ratio for transfemoral and transtibial BAPs was approximately AUD$17,000 and AUD$12,000, respectively, per quality-adjusted life-year compared to socket prostheses. This project was challenged by 17 barriers (e.g., limited resources, inconsistency of care pathways, design of preliminary cost-utility analyses) but eased by 18 facilitators (e.g., action research plan, customized database, use of free repositories). In conclusion, we concluded that lower limb BAP might be an acceptable alternative to socket prostheses from an Australian government prosthetic care perspective. Hopefully, this work will inform promoters of prosthetic innovations committed to making bionic solutions widely accessible to a growing population of individuals suffering from limb loss worldwide.


ervice (QALS)
a Queensland Health organization delivering artificial limbs to individuals suffering from limb loss.My principal mandate as manager of state service is to support the best possible prosthetic care while ensuring accountability for the use of taxpayer dollars.


Initial awareness

I became aware that osseointegration could provide opportunities for direct skeletal prosth

ic attachment in t
e early 1990s, when the first cases were presented at various international conferences by Dr. Rickard Branemark, a leading surgeon from the Sahlgrenska University Hospital, Gothenburg, Sweden. 1-3Similar to the rest of the prosthetic care community, I recognized the potential capacity of this surgical procedure to alleviate caveats of socket-suspended prostheses (SSPs). 4However, it was unclear how contraindications for consumers experiencing vascular problems and the inevitable adverse events (e.g., infections) that could lead to removal of the implant and reamputation should be dealt with. 5egardless, it was clear that the progress of this new treatment was remarkable and truly worth monitoring.Osseointe

ation was s
stematically included in QALS' regular horizon scans of prosthetic care innovations having potential to alleviate the clinical and financial burdens of prosthetic attachment for Queenslanders (e.g., review of literature about efficacy and safety).

A handful of patients were first fitted with a screw-type implant in 2000 by a team in Melbourne, Victoria, in collaboration with the pioneering group in Sweden. 6It also highlighted that bone-anchored prostheses (BAPs) could lessen expenditure from socket fittings and residuumrelated skin treatments. 4This was the first time I wondered how

he eme
gence of new treatments relying on direct skeletal attachment and the subsequent provision of BAP could impact the day-to-day work of a governmental organization such as QALS.


Challenges

Answering this question became critical when the first Queenslanders with unilateral transfemoral amputation were treated interstate in late 2012.Initially, we dealt with these consumers on a case-by-case basis.This approach was required to understand nd address immediate needs.However, it created too much uncertainty and unpredictability to be sustainable.Furthermore, we anticipated a significant influx of consumers in the short term.Soon after, QALS faced the challenge of putting in place a procedure to warrant a fair and equitable delivery of lower limb BAP to its consumers.


Needs

As n administrator, and often gatekeeper of taxpayers' money, I considered it essential to make decisions about a new treatment based on the best clinical and socioeconomic evidence available.Prosthetic care must be supported but resources are limited.Like many o her managers of government healthcare organizations, every dollar spent by QALS must be spent according to "financial marching orders" (e.g., schedule of allowable expenses).

Literature searches conducted during horizon scans and discussions with colleagues revealed that there was limited information about the alleged socioeconomic advantages of BAP. 10, 11 Clearly, there was a knowledge gap: What could the provision of BAP mea

for gover
ment healthcare organizations in terms of service delivery and expenditures?

In 2013, I initiated what turned out to be a five-year project of research gathering evidence to support the provision of BAP from the QALS perspective (Figure 1).We assessed the areas of disruptions while trying to find ways to accommodate new expectations.This project examined changes related to the service delivery of BAP, incl ding the development of a policy upporting the provision of BAP, the role of prosthetists, adjustments of continuous improvement procedures, and consumers' quality of ife.This project also involved a health economic evaluation of transf moral and transtibial BAPs, including cost compari on and preliminary cost-utility analyses (CUAs), compared to SSP.


Purposes

This paper reflects on the QALS experien e gained during this research project.The main purpose was to summarize the outcomes from a bird's-eye view.We have shared the lessons learned during our journey through hands-on information that might be helpful for all BAP promoters, including end users and carers, providers of prosthetic solutions, and administrators of healthcare organizations, amongst others.

The specific objectives were to:

• Introduce some background information about QALS' jurisdiction to facilitate cross-com arison and transferability of our experience

• Outline the drivers that motivated this work


LIST OF ABBREVIATIONS

• Present an overview of the impacts and outcomes

• Share the selected barriers and facilitators met during this project separately, although they were intertwined and

• Suggest briefly future developments of this work alongside some calls to action to further promote innovations in the service delivery and economic evaluation of BAP Drivers, barriers, and facili

tors we
eemed within and beyond QALS' influence were highlighted so that other organizations could identify their internal strengths and possible external threats during the strategic planning of similar resea

h projects (e.g.,
strengths, weaknesses, opportunities, and threats analysis).

Supplementary materials to be published in a Data In Brief paper provided addit onal information about the QALS' jurisdiction, publications (e.g., distribution, breakdown of impacts, downloads worldwide), allowable hours for prosthetist's labor (e.g., phases of treatment, tasks), study cohorts (e.g., sample size, representativeness), and datasets considered to estimate costs (e.g., number of cla ms, prediction), as well as detailed descriptions of all barriers and facilitators.


JURISDICTION

QALS is in the jurisdiction of the Queensland State Government Minister of Health, one of the six states and three territories of Australia.The role of QALS is to ensure equitable provision and funding of external prosthetic components to eligible residents of Queensland.Eligible consumers must be registered with the QALS and (1) be eligible for definitive prosthetic funding support under the Queensland Government's "Artificial Limb Scheme" or (2) be eligible under the Rehabilitatio

Appliance Program
of the Department of Veteran Affairs.QALS has a yearly budget of AUD$5.4 million to provide prosthetic services to 3,600 active onsumers annually through a network of up to 10 individual prosthetists (e.g., CPO).Although Queensland has predominantly an urban population, QALS services consumers across the whole state.

Queensland has hot and humid weather for the most part of the year.These conditions make the typical SSP difficult to tolerate and increase the need for frequent socket fittings.Access to the closest point of care can be particularly critical for some consumers who might have to travel hundreds of kilometers to visit their prosthetist for socket and component fittings.Altogether, the prospect of socket-free prosthetic solutions c uld be particularly appealing for QALS consumers.Currently, QALS is looking after a case-mix of nearly 100 consumers using unilateral, bilateral, and quadrilateral BAPs, representing approximately 11% and 6% of the existing population using BAP which is estimated at 950 in Australia and 1,600 worldwide, respecti ely.The number of QALS consumers has increased steadily by up to 10 per year over the last three years, generating one of the largest growing populations worldwide.


Continuous quality improvement


Role of prosthetists


Policy for provision


DRIVERS

Beyond our initial genuine interest in the economic impacts of the provision of BAP, this research project was pragmatically motivated by a series of external and internal dr vers to the organization.


External drivers

As hinted at in the historical introduction, this project emerged because of external drivers, including, but not limited to, the following:


•

A growing number of consum

, QALS st
rted to experience a significant influx of existing and new consumers choosing direct skeletal attachments.

Projections estimated that the number of consumers choosing BAP will continue to increase n

iceably, possibly re
ching between 150 and 200 consumers by 2025.

• Prosthetists' concerns.This project was also required to adequately recognize the hours spent by prosthetic care providers looking after consumers with BAP that should be supported by QALS.In 2012, there were no items within the existing QALS' schedules of allowable hours that prosthetists could claim after they provided standard care to fit BAP (e.g., no set hours for a specific service).The pathways for the compensation of their services were unclear.Providers could potentially experience improper compensation for fitting the BAP and loss of rev

ues from socket fittings
Internal drivers

This research project was also needed from several QALS organizational standpoints, including, but not limited to, the need to:

• Apply evidence-based practice.Like other government organizations, QALS was required to provide evidence supporting decisions about reimbursement standards, particularly for the provision of new health technology innovations that could be costly and obsolete within five years. 12

• Manage stakeholders' expectations.Clarification of the processes for service delivery of BAP was required to manage expectations from QALS stakeholders, including consumers and prosthetic care providers.Consideration whether the proposed procedures had legal bearings might be irrelevant (e.g., unlikelihood of lawsuits).Regardless, QALS believed that outlying these processes should help mitigate potential misunderstandings and conflicts inherent to the implementation of a new and, possibl

cost-saving potential.Economic
evaluations were required to confirm and, more importantly, to quantify if taxpayers' money could be saved with the provision of BAP, reducing the costs of socket fittings.Understanding cost-saving was essential to facilitate implementation given the budget constraints.

• Assist strategic planning.One of the most critical drivers was to gather sufficient information to complete the QALS' five-year strategic business plan, including yearly budgets for the provision of prosthetic care stratified by case-mix, including those with BAP.It was anticipated that the outcomes of this project would assist QALS with predictable workflow, help manage resources, and ultimately pl

a realistic b
dget.

• Take leadership.Perhaps less pragmatic but equally important was QALS' aspiration to take a leadership role in the area of health economic research on prosthetic osseointegration solutions that was then overlook

.


IMPACT


Overview

T
e overall impact of the project was summarized by nine key indicators which are presented in Figure 2 reflecting the publication outputs, scientific recognitions, and international acknowledgments.


Publication outputs

To date, we have authored a series of 21 publications between 2015 and 2020 (e.g., Digital

ject Identifier, Internation
l Standard Book Number), including six (48%) original research papers, one (5%) dataset paper, three (14%) repository papers, nine (43%) abstracts in national and international conferences, and two (10%) scientific annual reports.  Onlyanuscripts published or in press were considered here.However, several manuscripts are currently in prepa ation for submission to health economics and prosthetic care journals as well as open access repositories (e.g., Data In Brief).


Scientific recognitions

The recognition of each publication was assessed using conventional bibliometrics and altmetrics, including the number of views, downloads, and citations extracted from research institutions' repositories, social network sites for scientists, publishers' websites, and citation databases.To date, these publications have accumulated approximately 13,600 views, 6,500 downloads, and 95 citations, as detailed in Table 1.Citations of the three papers were in the 46th, 71st, and 46th percentiles corresponding to average, good, and above-average attention scores compared to other papers of a similar age in all journals, according to PharmacoEconomics-Open, Journal of Prosthetics and Items Views (1) Download (1) Citations (1) (2) Published in Canadian Pr

thetics & Orthotics J
urnal, Journal of Prosthetics and Orthotics, PharmacoEconomics-Open, Prosthetics and Orthotics International, The AOPA Revie; (3) Published in Data In Brief; (4) Published in Mendeley; (5) Presented at International Society of Prosthetics and Orthotics; (6) Presented at Australasian Osseointegrated for Amputees Conference Orthotics, and Prosthetics and Orthotics International, respectively. 14,16
(#) (%) (#) (%) (#) (%) (#)(

International acknowledgments

Analyses of ePrint records indicated that publications were downloaded from approximately 70 countries, with 75% of the downloads made from Australia (32%), United States of America (30%), Canada (6%), United Kingdom of Great Britain and Northern Ireland (4%), and Ireland (3%).5][36][37][38][39][40] This work provided guidance when the Australian National Disability Insurance Scheme developed its funding model.


CONTRIBUTIONS

The actual developments of each topic of research progressed altogether and often organically, depending on opportunities and resources.Therefore, contributions are presented by topics rather than historical evolution.


Health service delivery

Our primary contribution was the development of a policy regulating the provision of BAP-specific prosthetic care.Effectively, these procedures organized a workflow meshing role for prosthetists, a quality improvement of specific procedures, and assessment of overall consumers' experience and quality of life.


Policy for provision of BAP

In 2012, information from health technology assessments of direct skeletal attachment that could help develop this policy was sparse.An initial version of this policy was published in 2017 (e.g., tasks, documents, costs), including possible obstacles and facilitators to implementation. 14An equitable provision of transfemoral BAP was based on seven processes involving fixed expenses during the treatment and five processes regulating ongoing prosthetic care expenses.The cornerstone of this policy was the allowance of 22 h toward prosthetist's labor to support delivery of BAP care costing up to AUD$3,300 per consumer.A pros

etist could spend 2.5 h (11%),
2.5 h (11%), 6.5 h (30%), and 10.5 h (48%) during the preoperative, surgical, fitting of light and definitive limb prostheses, and postoperative phases of the treatment, respectively.

This policy required adjustments related to the prosthetists' scope of practice, funding of prosthetic limbs during rehabilitation, and allocation of microprocessor-controlled prosthetic knees.


Role of prosthetists

 In the policy presented earlier, prosthetists could claim up to 22 h of labor including 4 h (18%), 2 h (9%), 14 h (64%), and 2 h (9%) to consult with the clinical team, evaluate functional outcomes, fit light and definitive prostheses, and report progress to stakeholders before and after the surgical implantation of the osseointegrated fixation, respectively. 14 summarized in Figure 3, Frossard et al. (2018) further detailed the crit cal roles prosthetists could play during the provision of BAP, including referral of consumers (e.g., discussing fitting options, elucidating surgical procedures, selecting the surgical team). 17The survey presented by Frossard et al. ( 2019) indicated that 25% of QALS consumers found information about the surgical procedure from a prosthetist. 20As expected, prosthetists should be responsible for usual fitting tas

(e.g., selection of compon
nts, alignment of prosthesis, prevention of falls).However, as reported in Clark (2021), prosthetists also play a key role in the prevention of load-related adverse events when fitting bone-anchored bionics prostheses. 45[48][49][50] Altogether, this study showed that the provision of BAP has the potential to be slightly outside the usual scope of practice of prosthetists. 51Training opportunities by qualified experts, guidelines from suppliers of implants, and formal recommendations from governing bodies about prosthetic care of consumers fitted with BAP and business management that could help reduce risks are sparse, or even missing, in some jurisdictions.Prosthetists may potentially be exposed to increased risks when treating BAP consumers. 17


Quality improvement procedure

The implementation of the QALS policy for the provision of transfemoral BAP has subsequently created a need for a continuous quality improvement (CQI) procedure seeking to enhance consumers' experience with the QALS process, supporting the provision of BAP.2018) presented a BAP-inclu ive CQI procedure. 17,32A redesign study led to this procedure to collect, analyze, and report the experience of 65 QALS consumers who delivered BAP-specific prosthetic care, as presented in Figure 4.The proposed CQI procedure required 1.3 h of prosth

ist labor or 6% of the 22 h
llowed for the whole procedure presented above, costing AUD$213 per episode of care.The time spent by a prosthetist, consumer, and QALS staff represented 24%, 24%, and 53% of the CQI procedure, respectively.The costs of labor for prosthetist and QALS staff represented 70% and 30% of the CQI procedure, respectively.
CPOJ Special SPECIAL ISSUE
This study demonstrated that government organizations can redesign a CQI procedure for comprehensive appraisal of the provision of prostheses that could be: inclusive of BAP, affordable and swift for prosthetists.Achieving a minimally disruptive BAP-inclusive CQI procedure can be facilitated by adaptation of a procedure already in place (e.g., use of routing questions to indicate if the survey is for SSP or BAP).


Consumer's quality of life

Another integral part of the CQI procedure was to assess consumers' experience with the overall provision of BAP and changes in their quality of life after implantation of an osseointegrated fixation.Frossard et al. (2019) presented the outcomes of a 25question ad hoc survey, including 7 (28%), 5 (20%), and 13 (52%) questions about "Osseointegration Surgery Details," "Pre-Osseointegration Surgery," and "Post-Surgery Osseointegration", respectively. 20A total of 12 out of th

65 eligible QALS consumers completed the survey, givi
g a return rate of 18%.All respondents were "happy" with their BAP and indicated that "it works as it should", including 91% of respondents satisfied with the componentry fitted to their BAP.Key figures of the respondents' experience with efficacy and safety of the procedure are provided in Table 2.More importantly, all respondents reported a level of satisfaction and quality of life above eight and seven out of 10 after surgical implantation of the osseointegrated fixation and fitting with BAP, respectively.These outcomes sugge

that QALS policy about the provision of BA
seemed to contribute favorably to overall consumer satisfaction.

Altogether, this work provided benchmark information that can educate the design of patients' experience surveys and clinical trials looking at the effec s of bionic solutions on consumers' quality of life (e.g., built-in governmental CQI procedure).


Health economic evaluations

The QALS policy was validated by economic evaluations.e purposely chose to perform preliminary CUAs, as detailed below, when discussing barriers and facilitators. An overview of our approach to collect, extract, and analyze estimates of costs and utilities is presented in Figu

5.Total
osts combined actual and typical costs extracted from financial records and allowable expense schedules, respectively.Baseline utilities were extracted from the literature, while incremental utilities were assumed.


Efficacy Safety

• Respondents wear their BAP on average 17±6 hours per day

• 91% of respondents said their BAP supported their lifestyle needs

• 58% of respondents experienced some infec of their percutaneous part post-surgery

• Respondents experience an episode of infections the exit point of their percutaneous part postsurgery for an average of 145±170 days


Preliminary cost-utility analysis of transfemoral BAP

Frossard et al. (2017) cross-compared historical costs for t

provision of SSP with the s
mulated costs for transfemoral BAP (e.g., labor, parts). 14Costs were extracted from QALS regulatory documentation according to functional levels (e.g., K-levels) and low-cost, budget, and high-cost options for knee and ankle units.The provision of a transfemoral BAP was 18% and 79% less than SSP for the prosthetist labor and attachment costs, respectively.BAP was more economical by AUD$18,200, AUD$7,000, and AUD$1,600 when fitted with low-cost, budget, and high-cost options, respectively, compared with SSP for the highest functional level (i.e., K4).
CPOJ Special SPECIAL ISSUE
The average cost for the provision of transfemoral BAPs was approximately 40% (AUD$13,562±

D$16,497) more than SSP, which can b
partially offset by an increase of 0.815 QALY.The provision of a transfemoral BAP was costsaving and cost-effective for 19% and 88% of the consumers, respectively.The indicative ICUR for the provision of a transfemoral BAP was approximately AUD$17,000 per QALY and significantly below the WTP (Figure 6).


Preliminary cost-utility analysis of transtibial BAP

Frossard et al. (2021) r ported a preliminary CUA for six QALS consumers using transtibial BAP (Table 3). 18The average cost for the provision of transtibial BAPs was approximately 20% (AUD$5,604 ± AUD$12,180) more costly than SSP, which can be offset by an increase of 0.489 QALY.The provision of a transtibial BAP was more expensive and cost-saving for 67% and 33% of the participants, respectively.The indicative ICUR for the provision of a transtibial BAP was AUD$12,000 per QALY and significantly below the WTP (Figure 6).


Early evidence of health economic benefits

These studies evealed that early engagements with suppliers of prosthetic components particularly suited for transfemoral BAP can strongly impact the overall costs (e.g., economical advanced knee and foot/ankle units).These studies also highlighted that suppliers of osseointegrated fixations can influence the outcomes of CUA as the cost of their percutaneous parts (e.g., connectors, protective device) could offset the costs of socket fittings.In all cases, the provision of both transfemoral and transtibial BAPs appeared to be acceptable alternatives to SSP from an Australian government prosthetic care perspective.


BARRIERS

An overview of the 17 main barriers encountered during this project is presented in Table 4.A total of 5 barriers (29%) were related to service delivery, 11 (65%) to economic evaluation, and 5 (29%) to project management.A total of 4 and 13 barriers were deemed unlikely (e.g., access to limited resources, dealing with multiple funding allowances, addressing ethics issues, accommodating new national Overview of the approach appl ed to conduct preliminary cost-utility analyses (CUA) of transfemoral and transtibial bone-anchored (BAP) compared to socket-suspended (SSP) prostheses providing incremental cost-utility ratios (ICUR) based on incremental costs expressed in monetary units and utilities expressed in quality-adjusted life-years (QALY) that were compared to willingness-to-pay threshold (WTP) for small series of plausible scenarios (e.g., base-case, worst-case, best-case) over a six-year time horizon from Queensland Artificial Limb Service (QALS) prosthetic care perspective.schemes) and likely (e.g., face inconsistency of care pathways, design preliminary CUA, predict timeline of publications) to be met by other government organizations such as QALS.Here, we have only detailed the core barriers that set in motion cause-and-effect reactions onto other obstacles.


Access to limited resources

As with most prosthetic care departments, resources to undertake a research project of developing evidence-based policy are sparse.Unfortunately, we were unable to collaborate with services specialized in health technology assessment within the Minister of Health.In 2016, we applied for two unsuccessful grants (e.g., Defense Health Foundation Grants for Medical Research, Australian Centre for Health Services Innovation -Implementation Grant).Supports from other services and funders were curtailed by their perception that the provision of BAP was "too niche."Alternatively, the project was to run with QALS and its partner resources (e.g., staff time, consultancy).


Face inconsistency of care pathways

Another root cause barrier was the unpredictability of BAP care pathways corresponding to the onset of a series of interventions made by specialists during the course of treatment.Generic descriptions of the surgical procedures and rehabilitation programs specific to either screw-type or press-fit implants published by teams overseas were available when we started. 1,2,,6,7,9,56-62

Additional ad hoc guidance for specific aspects were provided regu

rly by main teams in Au
tralia as their own procedure evolved organically from case to case.Sometimes information from various sources agreed, but they often contradicted themselves.Consumers in the same case-mix rarely followed comparable care plans.Practically, it was difficult to grasp "who was doing what and when" around the fitting of BAP.Uncertainty about the continuum of care across preoperative, surgical, and postoperative phases of the treatment created the following barriers:

• Sort out schedules for allowable expenses.The adequate allocation of allowable hours to support the provision of BAP-specific prosthetic care was initially    39 However, the loss of income could be compensated by fitting BAP with high-end components.

• Dealing with diversity of outcome measures.Accessing clinical outcomes with osseointegrated implants is critical for heal h economic evaluations (e.g., choice of utility).However, assessing benefits, let alone harms, of surgical treatment was beyond QALS' prerogatives.Alternatively, we had to rely on a limited number of outcomes extracted from external sources. Ultimately, we preferred health-related quality of life data measured by the standardized 36-Item Short Form Survey (SF36) as the primary outcomes to reflect benefits and, more particularly, utility of the treatment. 7,57,64

• Palliate limited standards of prosthetic care.Inconsistent care pathways and diversity in outcome measures, all combined, hinder the understanding of the cause-andeffect relationships between treatment options, benefits, and harms (e.g., two-stage for screw-type, single-stage for press-fit). 65,66This limited the emergence of reasonable standards for BAP-specific prosthetic care, let alone the best standards around fitting arrangements that could possibly maximize benefits and minimize exposure to risks (Figure 7).However, the evaluation framework raised our awareness about the links between the risks of adverse events and loading regimen depending on the fitting of components as well as daily usage of BAP.Clearly, the choice of components can play a critical role in reducing loadrelated harms susceptible to osseointegration and the long-term stability of the bone/implant coupling. 46,48- 50,67-69 Initially, only a small case series showed differences between loading profiles applied by different categories of components (e.g., basic and advanced knee units). 69We examined mechanically passive components with basic functions such as single-axis or polycentric hydraulic knees and multiaxial foot-ankle units.Finally, we ackn wledged that the fitting of the microprocessor-controlled knee (MPK) and energy-storing-and-return (ESAR) foot was required.This decision was based on the best evidence available and, more heavily, on the alleged capabilities of these components to increase stability (e.g., stance and swing control), ease of walking (e.g., high range of motion, mechanically powered push-off), attenuate excessive loading (e.g., auto-adaptive stance and swing phases), and reduce falls (e.g., automatic stumble recovery). 70ltimately, we opted to support the provision of a "budget option."This package combines a single-axis cadence-responsive k ee, shock absorption adapter, tube adapter, and a dynamic foot that are commonly provided to QALS consumers with the highest functional outcomes (e.g., K4).


Design preliminary CUA

Undertaking preliminary CUA came with a range of subsequent obstacles to overcome when choosing the constructs framing the analysis, including, but not limited to, the following:

• Choosing a relevant perspective.First, we had to choose the perspective of the CUA corresponding to t e point of view adopted when deciding which healthcare costs should be considered.In principle, a comprehensive analysis could include all surgical, medical, and prosthetic healthcare costs covered by taxpayers.However, Queensland State healthcare organizations are structured in such a way that whether the fitting of BAP affects medical costs has little impact on the QALS' resources.We were more concerned with the potential reduction in the

ost of prosth
tic care.Therefore, CUAs were only conducted from the perspective of government prosthetic care.

• Establishing a relevant time horizon.The second obstacle was to determine the relevant time ho izon corresponding to the time over which outcomes of the innovation should be evaluated.indicated that it is often unclear how time influences both the technical adequacy of cost-effectiveness analyses and their correspondence to the policy choices they seek to inform. 73Osseointegrated implants are permanent, and fittings of BAP are continuous.At first glance, it could make sense to perform a comprehensive CUA using Markov decision-analytic models to look at multiple scenarios over scal

le time horizons (e
g., years, decades, lifetime).2015) demonstrated that the approximation error is large with the long cycle length and that the short cycle cost-effectiveness analyses better approximates the continuous-time reality. 73Furthermore, the World Health Organization recommended the production of generic cost-effectiveness analyses focusing on resources that could realistically be reallocated over the time horizon of the analysis. 72These recommendations lead us to make a compromise of a six-year time horizon, allowing a reasonable predicti n of the costs over the components' life cycle (e.g., two cycles of three years for a foot, three cycles of two years for a knee). 16,18

• Estimate costs.E penses from the QALS financial system for the provision of SSP or BAP were unavailable when individuals became QALS consumers less than six years before the surgery or when surgery occurred less than six years before the end of the study.As detailed above, the total costs were estimated by blending actual and typical costs.A prediction variable corresponding to r

ative typical costs over
he total costs, expressed as a percentage of the six-year funding cycle, was created to specify the level of uncertainty of the cost estimates.A prediction f 0% and 100% indicated that the total costs were fully extracted from the schedule and financial records, respectively.The overall cost predictions were 48±20% and 46±22% for the provision of transfemoral (SSP, 42±32%; BAP, 55±27%) and transtibial (SSP, 43±40%; BAP, 49±12%) prostheses, respectively.

• Access utilities.In princi le, utility data may have been obtained from the Australian treating teams.However, this option turned out to be impractical (e.g., access limited by ethics, no state-based stratification of datasets) and potentially unreliable (e.g., no clinical trial registration).These issues were resolvable.However, we chose to consider the quality of life status published previously. 7,8Baseline QALY were extracted from SF36 datasets converted into QALY applying regr

sion model.[16, 18] We made
conservative assumptions to determine the incremental gain of QALY between the SSP and BAP fitting options.

• Estimate the weight of the assumptions.By definition, preliminary CUAs overlook comprehensive uncertainty and sensibility analyses.Therefore, understanding the impact of assumptions to estimate individual costs (e.g., creation of a schedule of allowable expenses, blending of actual and typical costs) and utilities (e.g., extraction of baseline from literature, assumptions for incremen al gain) on both ICURs for transfemoral and transtibial BAPs was limited.

The choice of the preliminary CUA turned into a facilitator over time.Shortcomings might limit the strength of the evidence of cost-utility.However, this decision was critical in delivering the project on budget, on time, and with added value.Furthermore, publications of the outcomes contributed to the conversation about the relevance and possibly the standardization of preliminary CUAs to assess prosthetic care innovations.


FACILITATORS

An overview of the 18 key facilitators is presented in Table 5.A total of 4 (22%) facilitators related to service delivery, 10 (56%) to economic evaluation, and 8 (44%) to project management.

A total of 10 facilitators might be specific to QALS (e.g., engage with local research teams, involve a critical number of consumers, access to financial data, customize databases, share datasets, use of free repositories).Eight fa ilitators could be transferable to other organizations (e.g., frame action research plan, choose preliminary CUA, adapt rather than create procedures, engage with social media, monitor impact).Next, we only detailed the facilitators deemed the most critical.


Customize database

QALS' preliminary CUAs were facilitated by a piece of software purposely designed to:

• Import historical data from 1,840 vouchers exported from QALS' financial system for CUA of transfemoral (i.e., 1,598 vouchers) and transtibial (i.e., 242 vouchers) BAPs. 16,18Code individual expenses from 4,014 claims to identify whether there were for transfemoral or transtibial prostheses, SPP or BAP, labor (e.g., fitting prosthesis) or parts (e.g., prosthetic knees and feet units), at

chment (e.g.
socket, connectors), or prosthesis. 16,18

• Compare aggregated costs for in ividuals and groups over the time horizon with SSP and BAP before and after surgical intervention, respectively. 14,16,18

• Create reports including tables and figures formatted for internal communication (e.g., quarterly budget, annual reports) and publications of papers (e.g., manuscript, supplement). 14,16,18is database gave us the flexibility to run queries on demand to present the most up-to-date analyses and outcomes (e.g., new individual expenses to improve predictions).


Use of free repositories

Like most government organizations, QALS must make the outputs of the project freely available to taxpayers in Australia and elsewhere, in a timely manner.

We made the point to share original research, datasets, and repository papers including supplements and spreadsheets
CPOJ Special SPECIAL ISSUE
as well as abstracts and scientific annual reports available free of cost to the public, either from publishers' websites, social network sites for scientists, and/or research institution repositories (Table 1).

Availability of publications increased visibility and built up credentials.Portals provided a means to monitor the impact (e.g., numbers of views and downloads).In the long run, we hope access to primary information will encourage collaboration with other promoters of BAP and facilitate secondary observational studies (e.g., analyses of causeeffect relationships between confounders and provision of BAP) and literature reviews and meta-analyses.40


Frame action research plan

Perhaps more transferable were the lessons learned from the first steps of action research.Studies started with the planning phase, including practical tasks to define the project (e.g., identify problems to solve, root cause analysis, define objectives, profile case-mix), determine the deliverables (e.g., review regulatory obligations, conduct stakeholder's analysis, determine reporting expectations), and review the literature.The following tasks were particularly helpful and transferable:

• Gathering a reference group or a "think tank" including experts in service delivery, health economics, data analysis, prosthetics and clinical care, biomechanics, and consumer representatives that could, altogether, inform QALS management about the relevance and feasibility of research proposals.

• Creating a stakeholder matrix to organize controllers, promoters, providers, and advocates who can influence the provision of BAP (Figure 8).The immediate benefit of this exercise was to identify as exhaustively as possible all local, interstate, national, and international stakeholders.This task also required to clearly define the "power" and "interest" of a stakeholder corresponding to its capacity to influence allocation of resources and to provide prosthetic and medical care, respectively.Th

e matrices
ere most helpful in engaging and managing communication with all stakeholders (e.g., seek funding, present at conferences).

• Profiling the case-mix involved in a study by presenting the distribution of consumers according to demographics (e.g., sex, age, height, weight, body mass index), amputation (e.g., time since first amputation and BAP, cause, level, number of amputations, length of residuum), and access to care (e.g., distance between residence to providers and QALS) characteristics.This information was essential to characterize potential confounders and their impact on the provision of BAP.For example, knowing the distance between a consumer's residence and the closest service provider is critical to determine how access to care across a wide state can affect the quality of care.This characterization became valuable when discussing outcomes and writing papers.


FUTURE WORK

Future research will be undertaken in a global environment characterized by:

• Stronger evidence

f efficacy and
afety.Since this project, the body of peer-reviewed literature focusing on rehabilitation, prosthetic fitting, efficacy, and safety has grown noticeably.Health-related quality of life tend to be reported with a small range of surveys easing cross-comparisons between studies. 7,8,57-64However, there are still no straightforward standardized ways to report harms. 48-50 Little is known about long-term outcomes (e.g., influence of aging issues).Altogether, it is difficult to ascertain whether direct s eletal prosthetic attachment relying on percutaneous osseointegrated implants will overcome the "decline effect" as described by Harris (2016).CPOJ Special

• The emergence of global ecosystem.We are also witnessing the formation of a global ecosystem including a set of orga