Associations between frailty, implant and outcomes after primary knee replacement

The University of Oxford requires Hospital Episode Statistics (HES), Patient Reported Outcome Measures (PROMs) and Civil Registration (mortality) data for the purpose of an investigation to determine associations between surgical and patient factors on outcome following primary knee replacement. The proposed work would be undertaken by individuals from the Big Health Data Group, of the Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, at the University of Oxford. The work requires the HES, PROMS and mortality data controlled by NHS Digital to be linked to National Joint Registry (NJR) data - for which HQIP is the data controller - and would be supplied by HQIP’s data processor, Northgate Information Solutions. All data accessible to the team at the University of Oxford will be pseudonymised. Northgate Information Solutions will have no access to NHS Digital data under this Agreement.

The work will investigate factors relevant in the delivery of knee arthroplasty care and consists of two work packages:

- Package 1 addresses surgical factors (implant choice and surgical strategy) important in knee arthroplasty outcomes.

- Package 2 will investigate the role of patient frailty in knee arthroplasty outcomes.

The outcomes of interest are similar for both work packages (patient reported outcome measures, quality of life measures, mortality, revisions, re-operations and health care resources usage), and would be addressable using the same data request.

This data will be used for work in the public interest and will add to the existing knowledge base for knee arthroplasty, to better inform patients, clinicians and commissioners.

Knee replacement is one of the most common surgical procedures in the UK, with more than 800,000 performed over the last 9 years. Whilst the majority of patients experience improvements in pain, function and quality of life, there remain a group (estimated at 10-25%) who are dissatisfied after their knee replacement. It is important to better understand which patients are likely to benefit, those who may not experience a positive outcome, and potential avenues to improve success after primary knee replacement. This knowledge would feed into the shared decision pathway to surgery, in which the potential benefit and risk of the procedure are core components.

The importance of ongoing investigation in the field is recognised. There is public interest from patients undergoing knee arthroplasty to greater understand the value of the surgical procedure. This interest has been formalised. For example, the James Lind Alliance Priority Setting Partnership - which is made up of patients, carers and clinicians - identified areas of ongoing interest for knee replacement. These include identifying factors to improve post-operative outcomes, deciding what implants result in optimum outcomes and determining which patients are more likely to do well after knee replacement. The National Joint Registry is currently responsible for monitoring knee replacements, and commonly collaborates with external groups to investigate outcomes of arthroplasty and relies on GDPR Article 6(1)(e) as the lawful basis to do this. Article 9(2)(i) will also apply to the research carried out under this Agreement as processing is necessary for reasons of public interest in the area of public health.

Knee replacement involves the implantation of artificial materials into the knee to replace damaged surfaces, with the aim of reducing pain and improving function. This proposal concerns the investigation of medical devices. Given the number of these devices used, and the demand for this surgery, it is important to understand the impact of such procedures as part of care delivery.

The work described in this Agreement would add to the knowledge base around primary knee replacement. This will differ from existing pieces of work in two ways:

Firstly, the use of patient frailty as the independent variable in the assessment of outcome after knee replacement is novel. Existing pieces of work that explore the influence of patient health on outcome have done so using specific co-morbidities or co-morbidity indices (for example the Charlson Co-morbidity Index and the American Society of Anaesthesiologists grading). Frailty is a more recently developed concept, and considers patient health status in a more holistic way (through the consideration of additional facets of health such as psychological and social needs). Systems to assess patient frailty have been developed, and been shown to be valuable predictors in determining mortality and post-operative complications in existing studies. The value of frailty as an assessment tool has also been recognised by NHS England, and its assessment in primary care is now mandatory, to identify opportunities to intervene and optimise patient care. Assessment of how frailty influences outcome after knee replacement has not yet been done in the UK using these national datasets.

Secondly, existing pieces of work in the field of joint arthroplasty often report only a particular type of outcome (for example revision rate, patient reported outcome measures or complication rates). Such an approach can make interpretation of outcomes after surgery difficult, in that ‘success’ of surgery depends on the outcome being used, and the criteria for success. In addition, the reporting of isolated outcomes from different cohorts makes subsequent comparisons more difficult. A major strength of this work (made possible by the linkage of a data held by the National Joint Registry and NHS Digital), is that for each aspect of knee replacement care being assessed, different categories of outcome (implant survival, patient reported outcome, medical/surgical complications) will be reported. Reporting multi-modal outcome in this way will give a more comprehensive assessment of primary knee replacement surgery than has been available to date.

The Clinical Trials and Research Governance Department of the University of Oxford have reviewed and approved both work packages. Section 251 approval has been granted for the flow of identifying data from the NJR to NHSD for the purposes of data linkage. Both work packages have been reviewed by the NJR/HQIP and approved. No patient identifying information will be accessible to the study team at the University of Oxford, with the data being received from NHS Digital being pseudonymised prior to release.

The data subjects are individuals who have undergone primary knee replacement, as recorded in the NJR, since 2003. To determine the outcomes of these individuals requires the combined dataset comprised of data from the NJR, HES Admitted Patient Care (APC), mortality and PROMS. The NJR dataset contains detailed information pertaining to the surgical care of the individual at the time of surgery (anaesthetic grade and body mass index of patient, seniority of operating surgeon, implant details, use of cement), as well as recorded instances of revision (where the knee prosthesis has components added, removed or exchanged) and the indication for this. The use of HES APC and mortality register data is required to characterise patient groups, and to identify instances of health care input not captured by the NJR (for example re-operations to the knee, complications such as stroke, myocardial infarction or venous thromboembolism, and to capture out of hospital deaths not recorded in HES). The PROMs dataset provides critical information on the outcome of surgery from the patient perspective (using validated joint specific outcome questionnaires and quality of life questionnaires).

The data request is for records pertaining to patients identified by the NJR from 2003 onwards. Such data is required to investigate the outcomes of interest (see work package descriptions), which can only be captured by long term follow-up. NJR and HES data fields will also be required to control for confounding variables such as age, gender, deprivation and ethnicity when determining associations between treatment and outcomes. There are no practicable alternatives to this data request, given the number of records involved and retrospective nature of the work. With regards to data minimization, Northgate Information Solutions will provide to NHS Digital a restricted list of identifying details for the eligible patients from the NJR dataset, for subsequent linkage to the NHS Digital controlled datasets. As such no patients beyond those receiving primary knee replacement as recorded in the NJR will be included. Further to this, the data fields requested from each data set are limited to those required for analysis (to characterise patient groups, to control for confounding variables, to determine outcomes).

The Big Health Data Group at University of Oxford is the sole data controller and will receive, process and analyse the data and subsequently publish the findings. The NJR is part of the national audit programme of the Healthcare Quality Improvement Partnership (HQIP) and is managed by Northgate Information Solutions – HQIP’s data processor for this purpose - which will provide a list of all knee replacement as recorded in the NJR since 2003.

The project has been instigated and is being undertaken by a substantive employee of the Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences at the University of Oxford. This individual has an employment contract with the University of Oxford and is a Royal College of Surgeons (RCS)/NJR Fellow which allows him to undertake work in his home institution but his salary funding is provided by the RCS. Therefore, RCS is providing funding for the person who will undertake the work but is not providing funding specifically for the purpose of this work and has not determined that it should be undertaken. The purpose of the investigation has been determined solely and entirely by employees of the University of Oxford. Objectives of the work are focused on informing knowledge in the field of knee replacement care delivery within the NHS to meet recognised needs. The purpose of the work has received external review by the NJR Sub-Committee as part of the approvals process to undertake the work, as it includes the use of NJR data. Such review ensures that the proposed work falls within the NJRs thematic areas of interest.

Assessment of how the data will be processed to meet the proposed purpose has been undertaken by the senior team members (University of Oxford employees), who have extensive experience in the field.

Work Package 1:

Work package 1 aims to inform on surgical factors and implant factors that may affect patient outcomes following primary knee replacement (these represent points 3, 4 and 17 identified by the James Lind Alliance Priority Setting Partnership as important questions for patients needing knee replacement for osteoarthritis). CAG support under 18/CAG/0144 has been obtained for this work package.

The surgical factor of interest in work package 1 is the choice of surgical strategy, with regards to how much of the knee to replace, if disease is isolated to one area. One strategy is to perform total knee replacement (TKR) for all patients, the other is to perform unicompartmental knee replacement (UKR) where possible. UKR has been shown to provide faster recovery, higher patient reported outcomes, lower complication rates and greater economic benefit, but higher revision rates, when compared to TKR (an evidence base made up from randomised controlled trials, large registries and multiple cohort studies). In clinical practice patients referred for orthopaedic treatment will demonstrate varying patterns of knee arthritis, but it is estimated that 25-45% of patients would be appropriate for unicompartmental knee replacement. Despite this the NJR recorded level of UKR use has remained at 7-10%, suggesting underutilisation, with concerns around high revision rates likely a driver. Prior work has compared the outcomes of TKR to UKR, or the outcomes of UKR between high- and low-volume centre. This work will investigate whether UKR or TKR based surgical strategies provide benefits to the patient group as a whole, and better reflect practice in the NHS. Case-mix adjustments will be performed prior to determining associations between surgical strategy and outcomes of interest.

The second component of work package 1 is an investigation of the influence of knee replacement design on outcomes after primary total knee replacement. The 2018 NJR Report lists 67 different knee replacement systems as used in primary knee replacement during 2017, comprised of multiple systems across different manufacturers. Knee replacement systems demonstrate particular design philosophies (e.g. cruciate retaining, posterior stabilised, medial pivot etc.), aimed at improving outcomes. However, it is unclear to date what difference such changes make. To this end this component of the work would assess outcomes between 4 over-arching design principles in knee replacement (medial-pivot, gender specific, high-flexion and cementation choices in primary knee replacement), following case-mix adjustment.

Outcomes of interest for both components of work package 1 are change in Oxford Knee Score (a validated joint specific patient reported outcome measure) at 6-months, satisfaction after primary total knee replacement at 6 months, revision rates at 5 and 9 years, adverse events and mortality at 90-days and 1-year. Regression analysis techniques will be used to determine associations between implant type and the outcomes of interest.

Work package 2:

Work package 2 focuses on the influence of the patient on their outcome following primary knee replacement. The majority of patients receiving primary knee replacement are over the age of 60, with some degree of co-morbidity. It is appreciated that some health conditions are associated with less positive outcomes following joint replacement. Prior work has investigated the role of specific co-morbidities or used grading systems (such as the American Society of Anaesthesiologists grade - ASA) or co-morbidity indices (such as the Charlson Co-morbidity Index). However, such methods do not consider the patient in holistic manner. The concept of frailty has been developed to provide a more global assessment of patient health, and considers not only physical disease, but also psychological and social aspects relevant to patient care. Frailty measures have been found to be predictive of mortality, hospitalisation and nursing home measures. Further, when used in the hospital setting, frailty has been shown to be as good as, or better than, grading systems such as ASA in predicting mortality or poor outcomes after surgery. The value of frailty as an assessment tool has been recognised by NHS England, with assessment now routinely performed in primary care to identify vulnerable individuals. Given the high level of demand for knee replacement, and an aging population with complex health care needs, improved understanding of the interactions between frailty and outcomes after knee replacement is of value in the shared decision-making process. Such knowledge is recognised to be of value by the James Lind Alliance Patient Priority Setting Partnership (points 3 and 11 for hip and knee replacement).

Work package 2 will investigate associations between patient frailty and post-operative outcomes after primary knee replacement. Frailty will be determined using a validated frailty index, using International Classification of Diseases codes contained with HES records. Associations with outcomes after knee replacement will then be determined, after case-mix adjustment, between patients classified as being fit, mildly frail, moderately frail or severely frail. CAG support under 18/CAG/0143 has been obtained for this work package.

Outcomes of interest are change in quality of life (as measured by the EQ-5D utility index) at 6 months, change in Oxford Knee Score at 6 months, revision rates at 5 and 9 years, adverse events and mortality at 90-days and 1-year. Regression analysis techniques will be used to determine associations between implant type and the outcomes of interest. A health economics analysis will be performed to determine how frailty affects the cost of care delivery for primary knee replacement. If frailty is found to be a more predictive measure of outcome than existing measures, it would represent a valuable addition to the decision-making process for knee replacement surgery. As frailty is now routinely assessed in primary care, such information could be included in referrals for speciality care. For this purpose, information on all HES episodes for eligible patients is required.

The data controller has considered that the data requested under this Agreement is the minimum amount necessary in order to carry out this research.

The data under this Agreement will be processed as follows:

1. On behalf of HQIP, Northgate Information Solutions will transfer the following identifying fields to NHS Digital from the National Joint Registry (NJR):

- NHS number

- Date of Birth

- Gender

- Postcode

- NJR unique identifier

The NJR unique identifier is provided for each eligible patient from the NJR for the purpose of linkage to NHS Digital held data-sets.

2. Northgate Information Solutions will transfer to the Big Health Data Group (BHDG) at University of Oxford, pseudonymised records of eligible individuals.

3. NHS Digital will link the cohort identifying fields with pseudonymised data from the following data-sets:

- HES APC

- Civil Registration (deaths)

- PROMS

4. NHS Digital will disseminate the pseudonymised data along with the unique NJR identifier to the BHDG at University of Oxford

5. On receipt of the NJR/HQIP and NHS Digital data sets, the BHDG will be responsible for linking the two data sources, using the NJR unique identifier only.

The resulting pseudonymised record-level data-set will be linked, stored and analysed within the BHDG's Data Security Protection Toolkit (DSPT) compliant environment. There will be no linkage to any publicly available data. The linked data set containing data from NHS Digital will not be accessed by the NJR/HQIP. Only substantive employees of the data controller will have access to the data under this agreement. These employees of the data controller have been appropriately trained in data protection and confidentiality.

There will be no attempt to re-identify individuals from this data-set, and results will be presented at aggregate level with small numbers suppressed in line with the HES Analysis Guide. Within the BHDG, data would be stored in a secure environment (data will be stored on an encrypted password protected drive and kept within a data safe).

Data processing will result in a number of outputs, as detailed below. It is not expected that any outputs will result in small numbers (defined as ≤5), given the size of the population. However, if this is found to occur for any output, HES analysis guidance will be followed, with suppression of such numbers (as detailed in 6.1 of HES analysis guide).

Dissemination of results/outputs:

The work will inform on key areas directly relevant to patient care in knee arthroplasty. The assessment of patient frailty is already recognised as important in primary care. As the majority of patients receiving knee replacement are over 60, with varying levels of co-morbidity, understanding the influence of frailty on outcomes after surgery would be immediately informative in the shared decision making process. The advantages and disadvantages of unicompartmental and total knee replacement have been discussed previously, this work is novel in that it will determine the outcomes for patient populations treated at centres with differing strategies for knee replacement. This has the potential to directly influence surgical decision-making and delivery of surgical care. Finally, the consideration of different knee replacement designs will expand on the annual NJR report, by reporting a wider range of outcomes. Such findings will inform on the performance of knee replacements in current use.

Reports of findings will be submitted for publication in relevant academic journals, for dissemination to clinicians and researchers. It is anticipated that there will be three major publications from the submitted data (addressing frailty, reconstructive strategy and implant design in primary knee replacement respectively). These publications would be submitted for publication within 18 months of data receipt, and publication within 24 months.

Findings will also be disseminated via presentation at orthopaedic conferences (for example those held by the British Association for Surgery of the Knee and British Orthopaedic Association) and may be in the form of oral or poster presentation. Such conferences represent an important method of disseminating new findings to clinicians and researchers. Such findings can be presented in advance of formal publication in academic journals, again with a target of submission to relevant meetings within 18 months of data receipt.

In addition to the above, a formal report of the work will be submitted to the National Joint Registry and Royal College of Surgeons of England, as part of the standard arrangements which allow an RCS/NJR Fellow to support the work. Neither the RCS nor Northgate Information Solutions will have access to the data, nor a role in the analysis or interpretation. These reports can then subsequently be published on either the NJR or RCS websites, and included in annual reports (for example the RCS annual report of Fellows activities) made available on their public facing websites. These reports will be submitted at the end of 2019.

Communication of results/outputs:

Both the National Joint Registry and Royal College of Surgeons are key bodies in the communication of new results relevant to health care delivery in England (this includes both clinicians, healthcare managers and policy makers). As mentioned above, reports of the work will be submitted to both the NJR and RCS at the end of the first year after data receipt.

As the work will inform knowledge in areas identified as important to patients (as detailed in the James Lind Alliance Patient Priority Setting Partnership for Hip & Knee Osteoarthritis), the findings of the work will be discussed with the NJR, for consideration of patient review through the NJR patient network.

In addition to the above, results of the work will be added to the Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences (NDORMS) web pages for the project (to be created as part of the CAG approval requirements). In addition, NDORMS manage the institutional social media platform, on which recent findings by groups are announced.

The need for further understanding of factors influencing outcome following knee replacement is well recognised, and is highlighted in the James Lind Alliance Patient Priority Setting Partnership for Hip and Knee Osteoarthritis.

Specifically the planned work would address the areas of need concerning modifiable factors pre-, peri- and post-operatively that can influence outcome; identifying pre-operative predictors of success; identifying characteristics of individuals who benefit from knee replacement and those who do not and what is the best implant/prosthesis for best/safest outcomes.

If frailty is identified as a more specific predictor of outcome than current measures, it would demonstrate additional value to the now required assessment of frailty in primary care. Such information could then be included in referral pathways, and to improve the shared decision-making process for knee replacement.

Greater clarity as to the role of implant type would inform surgeons when it comes to selecting the implant for use, and to the NHS as a purchaser of implants. Similarly, an improved understanding of surgical strategy on both patient-oriented outcomes (quality of life, risks and complications) and health economics have the potential to support a wider change of practice in the NHS.

Not stated in the register.