Implementing patient-reported outcomes in clinical decision-making within knee and hip osteoarthritis: an explorative review

In the past few decades, there has been an increasing focus on the importance of patient involvement in the health care system. Patient participation executed through patient-reported outcomes (PROs) has been framed as being positive for the patients, care providers, as well as the health care system as a whole. [1, 2] PROs can be defined as measures of a patient’s state of health that come directly from the patient, without interpretation by others [3, 4]. Until recently, PROs have primarily been used for research purposes in clinical trials. However, interest is growing regarding integrating collection of PROs directly into clinical practice [5–8]. In this context, PROs hold a potential for facilitating shared decision-making by promoting patient involvement in their own treatment [5–8]. In addition, the integration of PROs into clinical practice can help to inform on indications for different treatment options and can be used to monitor the effectiveness of a treatment [7]. PROs are measured systematically using tools or instruments sometimes referred to as PRO measures (PROMs) [4], which can be of either a generic, disease-specific, or domain-specific character [4, 9]. Generic questionnaires, often measuring the quality of life (QoL) can be aggregated across different types of diseases. Disease-specific questionnaires focus mainly on a particular disease or condition, whereas domain-specific questionnaires aim at measuring specific domains of a disease, e.g. pain [4, 9].

A number of reviews have been conducted to date to evaluate the effects of using PROs as an integrated part of clinical treatment [7, 10–12]. These reviews generally have found insufficient evidence to support the implementation of PROs in clinical practice. However, to the authors’ knowledge, no systematic evaluations investigating the use of PROs in clinical practice within hip or knee osteoarthritis (OA) have been performed previously [10, 11]. First-line treatment of OA consists of nonsurgical treatment [13]. When OA is advanced and no other treatment is efficient, referral for surgical treatment is considered [13]. For patients with OA, information on PROs like pain and daily function is important for diagnostic and treatment considerations [14–16]. Furthermore, not fulfilling patient expectations has been found to be negatively correlated with satisfaction after total knee arthroplasty [17, 18]. Therefore, using PROs to improve shared decision-making and the matching of expectations might improve postsurgery patient satisfaction. As a result, using PRO as an integrated part of clinical practice can be highly relevant within hip and knee OA. Accordingly, in this review, it was aimed to investigate and discuss the current and future use of PROs in clinical practice and to identify the most common types of PROMs used for patients with hip or knee OA in different treatment settings.

The reporting of this explorative review follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, where applicable [19].

Integrating PROs into clinical practice has previously been proposed as a potential way to improve shared decision-making and patients’ involvement in their own treatment. However, there is currently a very limited amount of evidence regarding the active use of PROs within hip or knee OA. In this explorative review, only two pilot/feasibility studies reported on the integration of PROs in clinical practice to improve patient outcomes. PROs were primarily reported in studies regarding clinical effect, followed by studies on prediction models and validation studies. In 349 studies, a total of 38 different PROMs relevant for patients with either hip or knee OA were identified. The EQ-5D, WOMAC, and VAS questionnaires were the most commonly reported generic, disease-specific, and domain-specific PROMs, respectively. However, a large variation in the use of PROMs both within and between surgical and nonsurgical settings was found.

In this investigation, the literature search performed was broad, with the aim of including all available studies found concerning PRO within hip and knee OA. Due to the explorative nature of this review, a formal assessment of the quality of the included studies was not conducted. However, some limitations of the review process exist. Even though several databases of scientific literature were searched for this review, no search of the grey literature was conducted. It is acknowledged that PROs are applied within quality improvement efforts and in the maintenance of health care services [22] and some studies could therefore have been missed. However, the aim of the current review was to identify scientific research studies on the use of PROs, which is why the omission of grey literature is considered of limited importance. Also, for the majority of the studies, only abstracts and not full-text studies were reviewed, leading to a possible bias regarding the inclusion and classification of studies, as relevant information could have been overlooked. At last, the screening of most studies was performed by separating the total number of studies in two halves, with each half screened by one reviewer (two reviewers total). As a result, there is a risk that the two reviewers made or would have made different decisions from one another regarding the eligibility and classification of the studies they reviewed. However, to limit the risk of bias, both reviewers participated in a preliminary screening of the first 100 titles and abstracts and compared and discussed the decisions of eligibility and, from this, a high level of agreement (95%) was achieved. Furthermore, a random sample of 20% of the studies considered eligible for inclusion during the screening of title and abstract (70 studies) were assessed in full text by the four authors in two pairs of two. The authors agreed that none of the 70 studies in the secondary screening concerned PROs as an integrated part of clinical practice, thereby increasing the confidence in the categorisation of the studies.

In this explorative review, only two studies reporting on the active use of PROs as an integrated part of clinical practice within hip and knee OA were identified. Gakhal et al. [20] reported on the effects of continuous feedback provided to doctors about their patients’ pain and functional level. Separately, Slover et al. [21] described the integration of a uniform, web-based software system at two different treatment sites to create a multisite registry of secure PRO data that are usable in real time by both patients and physicians. Over the last 10 years, a number of studies [10, 11], have examined the use and application of PROs as an integrated part of clinical practice within different clinical areas and with varying conclusions. Importantly, the integration of PROs into clinical practice can be complex and resource-intensive [7, 10, 23]. As a result, Porter et al. developed a framework in which six key issues were highlighted. When considering to integrate PRO into clinical practice it is suggested to clarify the clinical activity that PROs are aimed at improving; to consider the characteristics of the PROM, the feedback system, the setting, and the target groups; and to support the implementation of PRO [7] Considering all of these requirements, it is not surprising that the integration and research into the effectiveness of PROs as a part of clinical practice can be challenging and that the results of interventions can and have differed. Based on the existing inconclusive evidence and the limited number of identified studies within hip and knee OA, more research on the application of PROs in clinical practice is advised to clarify whether integrating PROs into clinical practice works, for whom it works, and under which specific circumstances it works.

Previous systematic reviews, [24–26], have reported on the trends and use of different PROMs within total joint arthroplasty, but, to the authors’ knowledge, no other studies have examined the application of PROMs on patients with hip or knee OA across different settings or the use of PROs integrated in clinical practice. In accordance with the findings in the present study, previous systematic reviews have reported on a larger number of available PROMs within joint arthroplasty. One such systematic review by Ramkumar et al. [24] reported on the psychometric properties and variation among available PROMs within total knee arthroplasty. A total of 47 different PROMs were identified; however, it was concluded that no single ideal PROM was available. This finding was supported by Gagnier et al. [25], who identified 32 PROMs but concluded that many PROMs have limited evidence regarding their psychometric properties. Furthermore, in the present review, a large variation in the use of PROMs, both between and within settings, was identified. Siljander et al. [26] had the objective of analysing trends in the reporting of PROMs within total joint arthroplasty in four major orthopaedic journals. The study identified 42 different PROMs and found that the choice of PROM varied depending on the specific journal, time, treatment procedure, and geographic region. These findings support a hypothesis that suggests that the choice of PROM is currently arbitrary and dependent on traditions rather than based on standards and evidence of the best measuring properties. As a consequence, further exploration of PROMs to generate future standardisation is incited.

In addition to the suggested generation of standardisation within a surgical setting, the authors of the present review suggest that the possibility of standardisation across both surgical and nonsurgical settings should be investigated. The apparent advantages of standardisation across sectorial boarders would be to improve the communication ongoing between different practitioners and to optimise patient flow across sectors. In addition, a cross-sectorial PROM could be beneficial in the use of PROs as an integrated part of the treatment of hip or knee OA. However, it is acknowledged that various practitioners might have different needs and demands for information to provide the best possible treatment to the patient and that the importance of involving practitioners in the development of a cross-sectorial tool must be emphasised.

The aim of this explorative review was to investigate and discuss the current and future use of PROs in clinical practice and to identify the most common types of PROMs used for patients with hip or knee OA in different treatment settings. It is now known that limited evidence on the use of PROs as an integrated part of clinical practice for patients with hip or knee OA exist. It is suggested that further primary research will be necessary to clarify the effectiveness of integrating PROs into clinical practice within hip or knee OA before a systematic review with a risk of bias appraisal of the included studies will be relevant. Furthermore, it was found that the EQ-5D, WOMAC and VAS questionnaires were the most commonly reported generic, disease-specific, and domain-specific PROMs, respectively. However, it was also found that many different PROMs are utilised in studies concerning knee or hip OA. This suggests that there is limited agreement on a joint standard for the use of PROMs within hip or knee OA both within and across sectorial borders. The choice of PROM might be arbitrary and dependent on traditions rather than on standards and evidence of the best measuring properties. It is recommended that further research should investigate the possibilities for future standardisation of PROMs within hip or knee OA.