The patients evaluated in the model presented with pain characterized as intense, unpredictable, and emotionally draining resulting in the avoidance of physical activity.(24) In other words, the pain was severe enough such that there was an avoidance of physical therapy, which has consistently been shown to improve function.(25) This type of pain is also distressing enough that it affects the person’s quality of life, limiting activities of daily living and other recreational activities.(16) Additionally, these patients were < 65 years of age (i.e. non-Medicare).
The clinical guidelines used for treatment of knee OA in this assessment were the 2019 ACR/Arthritis Foundation guidelines for the management of OA of the hand, hip, and knee.(8) All therapies recommended by the ACR were used in the model. However, one therapy that was not recommended by the ACR was also used, hyaluronic acid (HA) injections (ACR has a conditionally recommended against for HA).(8) Inclusion of HA use was due to the identification of a 2009 Cochrane Review on viscosupplementation where HA was supportive of its use for relieving pain.(25)
Identification of the highest quality of evidence was used in the evaluation of outcomes for pain relief and function for all therapies. The types of studies identified therefore were systematic reviews and meta-analyses. PubMed and Cochrane Library were the databases searched with the key words used for identifying studies being:
((((((((((((((systematic) AND review) AND meta-analysis) AND pain) AND knee) AND osteoarthritis) AND outcome)) AND function)) AND random*) AND control*) AND trial)) AND NSAID – hits 292
((((((((((((((((systematic) AND review) AND meta-analysis) AND pain) AND knee) AND osteoarthritis) AND outcome)) AND function)) AND random*) AND control*) AND trial)))) AND physical therapy – hits 1,174.
((((((((((((((((((systematic) AND review) AND meta-analysis) AND pain) AND knee) AND osteoarthritis) AND outcome)) AND function)) AND random*) AND control*) AND trial)))))) AND hyaluronic acid – hits 186
(((((((((((((((((((((((systematic) AND review) AND meta-analysis) AND pain) AND knee) AND osteoarthritis) AND outcome)) AND function)) AND random*) AND control*) AND trial)))))))))) AND steroid) AND injection* - hits 211
((((((((((((((((((((systematic) AND review) AND meta-analysis) AND pain) AND knee) AND osteoarthritis) AND outcome)) AND function)) AND random*) AND control*) AND trial)))))))) AND ultrasound – hits 453. Figure 1 identifies the process of identifying the systematic review and meta-analyses used in the current analysis.
In order to simplify the model and not introduce confounding therapies, the assumption was made that therapies were used sequentially for both SAM treatment and SOC based on ACR care guidelines. In other words, if a therapy improved upon pain enough (with pain being the limiting factor for engaging in PT) such that PT could be initiated, it ended the decision to use another pain therapy. However, if pain was not improved, another modality was employed. The probability of improvement in pain such that PT could be initiated was identified in the systematic reviews and meta-analyses identified in the searches.(25–28)
Costs were derived from manufacturer pricing,(29) reimbursement rates over last 3-years for the SAM device including federal agencies TRICARE Military and VA Health System,(30) and from 2020 National Medicare rates adjusted by the ratio of Medicare payments to commercial payments in order to obtain the commercial rates.(31, 32)
Lastly, the statistically significant effect sizes for a therapy in improving pain relief and function were identified in the systematic reviews and meta-analyses. (27, 33, 34) The effect size is a simple way of quantifying the size of a difference between 2 groups. An effect size of greater than 0.5 generally means that the difference is important and can be observable which was used in the model.(35) It was further assumed that when PT was initiated (based on pain improvement), it was continued for the duration of the 6-month analysis as PT has been shown to have a positive effect on function over time.(36) The number of sessions and duration (e.g. dosing) was derived from the medical literature whereby the largest effect could be identified.(36)
Pain and function were established endpoints of the model as they are arguably of the most important outcomes in patients with knee OA. The American Academy of Orthopedic Surgeons (AAOS) has stated that: “the quality and success of interventions to treat OA should be assessed based on outcomes deemed to be of importance to the patient.”(37) The measure of pain and function in knee OA has also been established as part of the physician quality reporting system (PQRS #109).(38)
The incremental cost effectiveness ratio (ICER) was used in measure improvement of pain and function, and has been used in the past in knee OA for determining the ICER of a therapy.(39) Pain and function have been recognized as important outcome measures by the Institute of Clinical Economic Review, a leading organization in the area of value analysis of therapies.(40) For this particular analysis, the ICER was calculated as follows: (cost of SAM treatment minus cost of SOC)/(functional effectiveness of SAM minus functional effectiveness of SOC).
Tree Age Pro Healthcare 2020 software (TreeAge Software Inc., Williamstown, MA, USA) was used in the analysis. The variables, distributions and equations used in the model are found in Appendix 1 and 2. The decision tree is found in Fig. 2. Additionally, a tornado plot was employed to determine which variables in the model had the greatest effect on costs such that one therapy was found to be less expensive than the other. These variables were then examined individually in sensitivity analysis to identify the value at which one therapy became more expensive relative to the other and whether this value was realistic in everyday practice. An ICER scatterplot was also evaluated whereby the variables in the model were randomly varied 1,000 times using the low and high values as identified in the variables/distributions (Appendix 1).
Lastly included in the model, and as mentioned above, a wearable, long duration, continuous home-use ultrasound system that is self-administered, and capable of delivering 18,720 joules of ultrasound therapy for up to 4 hours/day generating vigorous deep musculoskeletal diathermy Δ8ºC at 1 cm, Δ6ºC at 3 cm, Δ3ºC at 5 cm; and; allows a patient to utilize ultrasound therapy outside of a medical facility (Sustained Acoustic Medicine [SAM] device, FDA Approval # K191568, ZetrOZ Systems, LLC, Trumbull, CT, USA). This device has been used safely and effectively in the treatment of knee OA,(20–23) chronic myofascial pain,(41–43) tendinopathies44–47 and for healing various musculoskeletal injuries.(19, 22, 44–47) Markov modeling (1,000 iterations) was used to examine on a percentage basis were SAM treatment would be a more effective and more costly therapy versus SOC.