If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Address correspondence and reprint requests to: N. Bansback, The University of British Columbia, 2206 East Mall, Vancouver, BC, V6T 1Z3, Canada. Tel.: 1-604–827–5453.
Shared decision-making supported by patient decisions aids may improve care and reduce healthcare costs for persons considering total joint replacement. Observational studies and randomized controlled trials (RCTs) have evaluated the short-term impact of decision aids on uptake of surgery and costs, however the long-term effects are unclear. This analysis aimed to evaluate the effect of patient decision aids on 1) use of joint replacement up to 7-years of follow-up, and 2) osteoarthritis-related health system costs.
Methods
324 participants in a Canadian RCT with 2-years follow-up who were randomized to either a decision aid (n = 161) or usual care (n = 163) had their trial and health administrative data linked. The proportion undergoing surgery up to 7-years were compared using cumulative incidence plots and competing risk regression. Mean per-patient costs were compared using two sample t-tests.
Results
At 2-years, 119 of 161 (73.9%) patients in the decision aid arm and 129 of 163 (79.1%) patients in the usual care arm had surgery. Between two and 7-years, 17 additional patients in both the decision aid (of 42, 40.4%) and usual care (of 34, 50.0%) arms underwent surgery. At 7-years, patients exposed to decision aids had a similar likelihood of undergoing surgery (HR = 0.92, 95% CI:0.73 to 1.17, p = 0.49) and mean per-patient costs ($21,965 vs $23,681, incremental cost: -$1,717, 95% CI:-$5,631 to $2,198) compared to those in usual care.
Conclusions
This is the first study to assess the long-term impact of decision aids on use of joint replacement and healthcare costs. These results are not conclusive but can inform future trial design.
Clinical trial registration
The full trial protocol is available at ClinicalTrials.Gov (NCT00911638).
While joint replacements are highly effective in relieving pain and improving function for most patients, up to 30% of recipients report insignificant symptom improvement and/or dissatisfaction with results
. Shared decision-making (SDM) supported by patient decision aids may improve the quality of care by helping patients set realistic expectations of surgery
, as evidence from trials of decision aids in elective surgery suggests that patients exposed to decision aids are more likely to choose conservative treatment.
The impact of decision aids on joint replacement rates varies depending on the context in which they are delivered. A 2017 RCT by Ibrahim et al. found that decision aids increased uptake of knee replacement in black patients (14.9% vs 7.7%, p = 0.04), who are less likely to undergo surgery than whites
. In populations with higher baseline rates of surgery decision aids may decrease uptake. A large observational study by Arterburn et al. found that the use of patient decision aids were associated with reductions in the use of hip and knee replacement (26% and 38%, respectively) and costs (12–21%) over 6-months
. A subsequent RCT found that decision aids for individuals considering hip and knee replacement were highly cost-effective, resulting in comparable outcomes and lower costs at 2-years follow-up
. Cost reductions were driven primarily by fewer surgeries in those exposed to decision aids, but importantly this trial was not powered on this outcome (73.2% vs 80.5%; Relative Risk = 0.91, 95% CI: 0.81 to 1.03).
Given the progressive nature of osteoarthritis, the long-term impact of decision aids on use of surgery is unclear, and this has been raised as an area for future research
. For instance, Arterburn et al. stated “it is entirely possible, given the natural history of osteoarthritis, that patients who choose to forgo joint replacement will reverse their decision later.”
Delaying surgery can be beneficial as it delays the cost of surgery and may reduce the likelihood of needing a subsequent revision. However, delaying joint replacement may result in a more complicated or costly surgery later, or increase the use of other health care resources, such as physician visits or medications to manage pain. The purpose of this study was to link RCT and health administrative data to evaluate the effect of decision aids on 1) use of joint replacement up to 7-years follow-up, and 2) osteoarthritis-related health system costs.
Methods
Detailed methods and results from the Canadian two-site, parallel group, single-blind, two-arm RCT have been reported elsewhere
, and the protocol was registered (#NCT00911638). Briefly, 343 patients with moderate to severe hip or knee OA were recruited (May 2008 to October 2009) from two orthopedic screening clinics in Ottawa, Ontario, Canada. Eligibility was assessed using a seven-item priority tool completed by physicians and developed using key guideline criteria including pain in motion and at rest, functional limitations (i.e., ability to walk), abnormal findings from physical examination, and the potential for disease progression based on radiographic evidence and clinical expertise. Each item was scored on a three- or four-point scale (e.g., ranging from mild to severe pain at rest)
. Participants were randomized to receive either a decision aid and one-page preference report for their surgeon (intervention) or usual care. A statistician constructed the randomization schedule, which was a centrally located and computer-generated block randomization procedure (randomly varying lengths of 4, 6, or 8 in a 1:1 allocation ratio which was stratified by site and joint). Participants were allocated using call-in telephone software. The decision aids (hip and knee) were developed by the Informed Medical Decisions Foundation and consisted of a DVD and booklet, while the preference report summarized the patients' knowledge, values, preferred treatment choice, and clinical assessment
. The primary outcome of the original trial was wait time (from eligibility assessment using the priority tool to reaching a decision on whether to undergo or forego surgery) and was powered to detect a mean difference of 8 weeks between the two arms, but was not powered on the outcomes assessed in this study (uptake of surgery and cost).
At enrollment, trial participants were asked to provide their personal health number and consent to having their trial data linked to administrative databases. Participants did not need to consent to this linkage to participate in the trial. In April 2017, trial data were linked to de-identified data sets using unique coded identifiers at the Institute for Clinical and Evaluative Sciences (Ottawa, Ontario). Administrative data sets included basic demographic information and vital statistics (Registered Persons Database (RPDB)), hospital discharge abstracts and same day surgeries (Discharge Abstract Database (DAD)), emergency department admissions (National Ambulatory Care Reporting System (NACRS)), physician billings (Ontario Health Insurance Plan (OHIP)), rehabilitation services (National Rehabilitation Services (NRS) database), and prescription medications for all individuals aged 65 and older (Ontario Drug Benefit (ODB) database). Canadian healthcare administrative databases are among the highest-quality in the world
Validity of administrative data in identifying complex surgical site infections from a population-based cohort after primary hip and knee arthroplasty in Alberta, Canada.
Exclusion of patients with sequential primary total joint arthroplasties from arthroplasty outcome studies biases outcome estimates: a retrospective cohort study.
Total hip and knee replacement were identified in the DAD using previously validated procedure and diagnostic codes from the International Classification of Disease, Canadian Classification of Health Interventions (ICD-10-CA/CCI)
Exclusion of patients with sequential primary total joint arthroplasties from arthroplasty outcome studies biases outcome estimates: a retrospective cohort study.
. Total joint replacements were identified using procedure codes starting with “1VA53” for hip replacement, and “1VG53” for knee replacement. Location of the procedure (i.e., left, right, bilateral) was identified using a supplementary status attribute, “inatloc,” corresponding to “L,” “R,” and “B.” Revision surgeries were identified using the “incode” status attribute, where “R” indicated the surgery was a revision. The proportion of patients undergoing joint replacement were represented using cumulative incidence plots that accounted for competing risk of death
. Risk of surgery over the follow-up period was estimated using competing risk regression (accounting for death) which controlled for joint (knee vs hip) and clinic site
. Deaths were identified by the “dthdate” field from the RPDB.
Health care system costs
Resource use and costing was conducted from a health system perspective and all osteoarthritis-related activities were included in the base case analysis. This includes all surgeries (index, revisions, and secondary surgeries on non-index joints), and primary care, specialist, and hospital-based physiotherapy visits. The DAD, NACRS, and NRS databases were searched with relevant ICD-10 codes to identify surgeries and admissions for OA or complications related to joint replacement (e.g., deep vein thrombosis). The OHIP database was searched to identify physician billings for joint replacements, in addition to any additional billings with a primary diagnosis of OA. OA-related drugs were identified in the ODB by Anatomical Therapeutic Chemical (ATC) category. All drugs in ATC categories ‘M’ (Musculoskeletal) and N02 (Nervous system – analgesics) were identified using their drug identification number.
Costing followed guidelines on person-level costing using administrative databases in Ontario developed by the Health System Performance Research Network
. For example, the cost of resource use identified in the DAD, NACRS, and NRS was estimated by multiplying a resource weight (e.g., ‘resource intensity weight’ for the DAD) by an average cost per weighted case. Cost of physician and laboratory services (OHIP) and prescription drugs (ODB) were taken directly from their respective databases. All costs were adjusted to 2016 Canadian dollars and discounted at 1.5% per Canadian guidelines.
Mean per-patient costs were compared using Welch two sample t-tests. Subgroup analysis was undertaken to evaluate the impact of the intervention on mean per patient costs for those considering knee replacement, and sensitivity analyses evaluated the impact of censoring resource utilization and costs at: 1) a second primary surgery (i.e., hip or knee replacement regardless of their initial surgery), and 2) at a second primary surgery on a different joint (e.g., hip replacement if their initial surgery was a knee replacement). These two sensitivity analyses were completed for all participants and the subgroup considering knee replacement.
Results
Of the 343 trial participants, 324 provided a personal health number which allowed their trial and administrative data to be linked (Table I). Of the 19 who were not linked, 13 were from the intervention arm while six were from the usual care arm. On average, patients in the intervention arm were followed up for 6.8 years (SD = 1.1 years) compared to 6.7 years (SD = 1.0 years) in the usual care arm. Overall, a greater proportion of participants were considering knee replacement compared to hip replacement (236 vs 88).
Table ISample characteristics
Decision aid arm (n = 161)
Usual care arm (n = 163)
Age (yrs), mean (SD)
66.1 (9.8)
67.0 (9.9)
Joint (n)
Hip
45
43
Knee
116
120
HKPT (total 80), mean (SD)
45.2 (13.7)
45.5 (13.4)
WOMAC (total 96), mean (SD)
56.0 (17.2)
52.9 (15.9)
Sex (n)
Men
77
62
Women
84
101
BMI, mean (SD)
30.8 (6.4)
31.7 (6.1)
Education (n)
< HS
10
12
HS/TS
74
69
College
31
24
University
28
39
Graduate School
18
19
Living arrangement (n)
Alone
36
42
With someone
125
121
Employment (n)
Full time
31
32
Part time
11
12
Retired
101
105
Other
18
14
Household income
<$20,000
14
10
to $39,999
25
34
to $59,999
38
35
to $79,999
33
22
to $99,999
15
16
>$100,000
27
31
no response
9
15
SD: standard deviation; HKPT: Hip Knee Priority Tool; WOMAC: western Ontario and McMaster universities osteoarthritis index; HS: high school; TS: trade school.
. Between two and 7-years, 17 additional patients in both intervention (of 42, 40.4%) and usual care (of 34, 50.0%) arms underwent index surgery, resulting in 136 intervention (84.5%) and 146 (89.6%) usual care patients having undergone index surgery over the 7-year follow-up. Competing risk regression estimated a hazard ratio of surgery of 0.92 for individuals exposed to decision aids (95% CI: 0.73 to 1.17). A similar estimate was observed in the subset of participants considering knee replacement (HR = 0.85, 95% CI: 0.64 to 1.12).
Fig. 1Cumulative incidence of surgery by treatment arm. UC: Usual Care Arm; DA: Decision Aid Arm.
Patient exposed to decision aids had similar mean per-patient costs compared to those in usual care ($21,965 vs $23,681, incremental cost: -$1,716, 95% CI: -$5,631 to $2,198) (see Table II). Analysis of participants only considering knee replacement found similar results in those exposed to decision aids ($21,043 vs $23,932, incremental cost: -$2,889, 95% CI: -$7,469 to $1,691), as did sensitivity analyses which considered censoring costs after secondary surgeries (Table III).
Table IIMean per-patient costs (2016 CAD$), by database
This is the first study to assess the long-term impact of decision aids on use of joint replacement and healthcare costs. While these results are not conclusive, they do address a gap in knowledge by evaluating the impact of decision aids on uptake of joint replacement over the longer-term. It is not clear why more patients did not have surgery over the 7-year period given the progressive nature of OA. They may have felt that the risks of surgery outweighed the benefits, or found that non-surgical treatments effectively managed their symptoms
. In relative terms, the potential reduction in use of joint replacement observed in this study is small, however given the number joint replacements performed annually this could translate into substantial costs avoided
. These resources could be re-allocated to populations who are less likely to undergo surgery as a means of reducing disparities in access and outcomes.
An important factor in contextualizing these findings is that use of joint replacement was not the primary outcome of the original clinical trial, and the trial was not powered to detect a difference on this outcome. The uncertainty in the evidence of the long-term impact of patient decision aids on treatment uptake and costs suggests a larger study may be warranted, and these results provide useful information to inform a future trial. For example, very few individuals undergo joint replacement after the first 5 years, suggesting that a trial with a 5-year time horizon may suffice.
Several limitations need to be considered. This study was conducted in Canada, in a context where surgeons are paid fee-for-service, patients have publicly funded healthcare, and included a population with moderate-to-severe OA after nearly half of screened patients were referred back to primary care with mild symptoms
. Thus, the results may not be generalizable to other contexts, such as jurisdictions where patients have less severe disease, or where physicians are salaried. In addition, this study did not consider patient outcomes, which were comparable at 2-years follow-up
, but could have changed over time. Poorer health could have manifested through increased use of resource utilization (e.g., physician visits or prescriptions), but we did not find evidence of this. Lastly this analysis did not capture prescription drug costs for individuals under 65 years of age or physiotherapy visits in non-hospital settings.
This study addresses a noted gap in the literature by elucidating the long-term impact of patient decision aids on use of joint replacement and costs. The results are not conclusive and future research is warranted. This includes understanding the differential impact of patient decision aids on use of joint replacement, costs, and outcomes in different populations and health system contexts over the long-term.
Author contributions
All listed authors: 1) contributed to the conception and design of the study, or acquisition of data, or analysis and interpretation of data, 2) contributed to drafting the article or revising it critically for important intellectual content, and 3) approved the final version to be submitted.
Role of the funding source
The study sponsors did not play a role in the study design, collection, analysis, or interpretation of the data, in writing the manuscript, or the decision to submit the manuscript for publication. This study was supported by the Institute for Clinical Evaluative Sciences (ICES), which is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). The opinions, results and conclusions reported in this paper are those of the authors and are independent from the funding sources. No endorsement by Institute for Clinical and Evaluative Sciences (ICES) or the Ontario MOHLTC is intended or should be inferred. Parts of this material are based on data and/or information compiled and provided by CIHI. However, the analyses, conclusions, opinions, and statements expressed in the material are those of the author(s), and not necessarily those of CIHI. The ICES is a prescribed entity under section 45 of Ontario's Personal Health Information Protection Act. Section 45 authorizes ICES to collect personal health information, without consent, for the purpose of analysis or compiling statistical information with respect to the management of, evaluation or monitoring of, the allocation of resources to or planning for all or part of the health system. Projects conducted under section 45, by definition, do not require review by a Research Ethics Board. This project was conducted under section 45, and approved by ICES′ Privacy and Compliance Office.
Conflict of interest
The authors have no conflicts of interest to declare.
Acknowledgments
This original study was supported by funding and access to the decision aids from the not-for-profit Informed Medical Decisions Foundation (Grant #0099-1)
Validity of administrative data in identifying complex surgical site infections from a population-based cohort after primary hip and knee arthroplasty in Alberta, Canada.
Exclusion of patients with sequential primary total joint arthroplasties from arthroplasty outcome studies biases outcome estimates: a retrospective cohort study.
30909-2&id=gr1.jpg){kind=link}