View Article PDF

Osteoarthritis (OA) is a common and debilitating chronic disease and one of the leading causes of disability in New Zealand and worldwide.1–3 Knee OA is the most common form of OA, affecting as much as one-third of the population at some point in their lifetime. Globally, OA is responsible for more than 16 million disability-adjusted life years (DALYs) and is the 12th leading cause of worldwide disability.4 In New Zealand, osteoarthritis accounted for more than 15,000 DALYs in 2018.5 In addition to the health losses associated with OA, it also accounts for a substantial economic burden on the health system and the wider economy. In New Zealand, the total annual financial cost of arthritis (of all types) is estimated to be $4.2 billion, including health system costs of $990 million; productivity costs of $1.2 billion, through reduced employment, time off work and impaired performance at work; and other costs, such as non-health sector care and aids and modifications to support independent living for people with arthritis, of $2 billion.5

Total knee replacement (TKR) surgery is a common and successful operation to reduce pain and improve HRQoL in patients with advanced knee OA.6,7 Provision of TKR has been increasing across the world in recent decades.8–11 However, TKR is a costly procedure and high provision rates place strain on limited public healthcare resources. Furthermore, access to TKR, particularly in the public healthcare system, is limited by capacity constraints due to the availability of a suitably-trained workforce and surgical facilities.

The two most important risk factors for incidence and progression of knee OA are age and obesity. In New Zealand, as elsewhere, the combination of an ageing population and increasing rates of obesity is therefore expected to result in continuing increases in demand for OA healthcare.8,10,12–15 The population aged over 65 is expected grow by 40 percent over the next decade,16 while obesity rates are projected to increase to close to 50% of the adult population by the late 2030s.17 Understanding the implications of these population changes is critical for successful health system and workforce planning.

The aims of this study are (1) to estimate the projected healthcare costs of osteoarthritis care in New Zealand for the period 2013–2038, (2) to estimate the demand for TKR surgery in New Zealand over the same period, and (3) to assess the contribution of projected increases in population obesity to future healthcare expenditure and TKR demand.

Methods

We used the New Zealand Management of Osteoarthritis (NZ-MOA) model, a state-transition microsimulation model of the incidence, progression, health impact and healthcare costs of knee OA in New Zealand.18 The model generates a hypothetical cohort drawn from the 2013 New Zealand population distribution; runs each simulated individual through a sequence of annual health state transitions, defined by obesity, knee OA status (Kellgren-Lawrence grade), health-related quality of life (HRQoL) outcomes and treatment pathway; and computes the resulting per-capita healthcare costs, quality-adjusted life years (QALYs) and treatment utilisation. By running the same cohort through multiple scenarios, the incremental change in outcomes attributable to knee OA prevalence, change in risk factors or change in treatment patterns can be estimated.

We computed model outcomes under two scenarios: continuation of projected trends in population obesity levels17 and holding population obesity constant at 2013 levels. We generated simulation cohorts of 10,000 individuals for each age- (in five-year age groups), gender- and ethnicity-specific population subgroup. For each subgroup, we ran each scenario for a modelled 25-year time horizon (2013–2038) and recorded the age-, gender- and ethnicity-specific OA-related healthcare expenditure and TKR incidence in each year. Projections of healthcare costs and joint replacement provision at the population level were calculated by multiplying the age-, gender- and ethnicity-specific per-person projected outcomes from the simulation model by Statistics New Zealand’s published national population projections for each future period.19 This whole process was repeated 1,000 times, each with a new random draw from the distribution of obesity projections, resulting in a total simulated population of 760 million individuals to provide stable and precise estimates of the outcomes. We calculated the mean healthcare costs and joint replacement incidence in each year across the 1,000 cycles to provide the projected point estimates for each outcome, and the 2.5th and 97.5th percentiles to construct 95% uncertainty intervals for the contribution of increasing obesity levels (ie, undertaking probabilistic sensitivity analysis).

Baseline population model input parameters and sources have been described previously,18 and are summarised in Appendix A. Costs of treatment were based on provision of usual publicly and privately provided medical care as practiced in New Zealand,20 valued with New Zealand-specific reference prices21 (see Appendix Table A5). TKR incidence rates in the New Zealand population were obtained from the New Zealand Joint Registry,22 which covers almost all privately and publicly funded joint replacement surgeries in New Zealand, and used to calibrate the baseline modelled TKR provision rates. Projected increases in population obesity rates, stratified by age, gender and ethnicity, were obtained from a previously published age-period-cohort model of body mass index (BMI) in the New Zealand population.17

Cross-model validation was conducted by comparing our projected increases in the provision of TKR with those previously published in the Journal by Gary Hooper and colleagues (2014).13 Those estimates were based on different inclusion criteria than our model: all TKR were included for any primary diagnosis, whereas our model includes only one OA knee per person (ie, only the first primary TKR performed), and includes only TKR performed for a primary diagnosis of knee OA. The total number of TKR performed annually should therefore be lower in our model projections; to allow direct cross-model comparison, we normalised all estimates to an index value of 100 in 2011 (the base year in Hooper et al’s projections), to compare the total percentage increase in TKR provision between models. (The observed increase in TKR provision between 2011 and 2013 was used to determine the baseline (2013) index value for the NZ-MOA model, as the model was calibrated to observed 2013 data.)

Results

In 2013 (the base year for the NZ-MOA model), the total direct healthcare costs of knee OA were estimated to be NZ$199 million (1NZD≈0.82USD), and 5,070 patients had a first total joint replacement for knee OA. Women accounted for $110 million (55%) of the total costs and 2,460 TKR (48%); Māori for $18 million (9%) of the total costs and 490 TKR (10%).

The healthcare costs associated with OA treatment were projected to increase to $370 million per year in 2038 (at constant 2013 prices; Figure 1), an increase of 85%. Per-capita treatment costs (for the adult population aged 25 and over) were projected to increase from $69 to $90. TKR incidence was projected to increase to 9,040 per year in 2038 (Figure 2). The provision rate was projected to increase from 174 to 221 per 100,000 population per year.

Figure 1: Projected healthcare costs of knee osteoarthritis in New Zealand, 2013–2038.

c

Source: NZ-MOA simulation model. Lines show the projected increase in annual healthcare expenditure on OA treatment under the base case scenario (ongoing increase in population BMI based on current trends) and with no further increase in population BMI (ie, maintaining 2013 BMI distribution), relative to 2013 baseline level. ‘Attributable to BMI increase’ is the difference between the two lines, representing the incremental OA-related healthcare expenditure attributable to continued increases in population BMI relative to the 2013 level. The shaded uncertainty interval represents the 95% confidence interval for projected annual increase in population BMI. Right-hand axis shows total projected annual healthcare expenditure in each period for the base case and no BMI increase scenarios.
BMI indicates Body Mass Index; OA, Knee osteoarthritis.

Figure 2: Projected demand for first total knee replacement for knee osteoarthritis in New Zealand, 2013–2038.

c

Source: NZ-MOA simulation model. Lines show the projected increase in annual provision of TKR under the base case scenario (ongoing increase in population BMI based on current trends) and with no further increase in population BMI (ie, maintaining 2013 BMI distribution), relative to 2013 baseline level. ‘Attributable to BMI increase’ is the difference between the two lines, representing the incremental TKR provision attributable to continued increases in population BMI relative to the 2013 level. The shaded uncertainty interval represents the 95% confidence interval for projected annual increase in population BMI. Right-hand axis shows total projected annual provision in each period for the base case and no BMI increase scenarios.
BMI indicates Body Mass Index; TKR, Total knee replacement surgery.

In 2038, women were projected to account for $210 million (57%) of the total healthcare costs and 4,760 TKR (53%); Māori for $44 million (12%) of the total costs and 1,120 TKR (12%). Full results for projected costs and TKR incidence for each year 2013–2038, including ethnic and gender breakdowns, are reported in Tables B1 and B2 in the Appendix.

Further increases in population BMI, relative to the baseline distribution in 2013, accounted for 25% of the projected increase in per-capita OA-related healthcare costs and 47% of the increase in TKR provision rates between 2013 and 2038. By 2038, this accounts for an additional $22 million and 880 TKR per year. Over the 25-year time horizon, projected increases in BMI, relative to baseline levels, will result in an additional $231 million in OA healthcare costs and 9,880 TKR required.

The increase in TKR provision projected by our model, in the absence of further increases in population obesity, was very close to that previously reported, using a different modelling approach, by Hooper et al (2014) (Figure 3).13 Between 2011 and 2026, TKR provision was projected to increase by 48.8%, compared to 49.0% reported in the previously-published estimates. Allowing for projected increases in population obesity, this increased to 56.6% in our model, and 90.8% by 2038. The total numbers of projected TKR (without normalisation) from both models are shown in Figure B1 in the Appendix.

Figure 3: Observed and projected demand for total knee replacement in New Zealand, 1999–2038, cross-model comparison.

c

Source: NZ-MOA simulation model; NZJR 2016;22 Hooper et al 2014.13 Points show observed provision of TKR 1999–2013; dashed line the published projections from Hooper et al 2011–2026; and solid lines the new model projections 2013–2038, under the base case scenario (ongoing increase in population BMI based on current trends) and with no further increase in population BMI (ie, maintaining 2013 BMI distribution). The absolute numbers in each series differ due to different inclusion criteria: all values are normalised to 2011 = 100 for consistent comparisons.
BMI indicates Body Mass Index, in kg/m2; NZJR, New Zealand Joint Registry; TKR, Total knee replacement surgery.

Discussion

Treatment of knee OA cost the New Zealand healthcare system $200 million in 2013, an amount projected to increase to $370 million over 25 years. Demand for TKR is projected to increase by almost 80% over the same period, requiring an additional 4,000 surgeries to be performed annually by 2038. After adjusting for population growth, projected increases in obesity rates account for one-quarter of the increase in per-capita costs and almost half of the increase in TKR provision rates over 25 years.

These results highlight the need for effective—and cost-effective—treatment of OA throughout the disease course. Our projections were calculated under the assumption that patients’ access to TKR (and other treatments) remains at the same level as in 2013. Access to TKR is already rationed through the public health system, with many local health systems unable to offer surgery to all patients who would benefit;23,24 improving access to TKR would require further resourcing beyond that suggested by our projections. Effective, low-cost, early interventions, such as exercise therapy, can alleviate symptoms, improve patients’ quality of life and reduce the need for costly treatment, such as TKR, later in the disease course.25,26 Improving access to such treatments may help to mitigate the increasing burden on the health system associated with rising rates of OA.

There is limited existing research on the future healthcare costs of OA. Our projected increases in costs are very similar to those reported recently for Australia,27 which had a projected increase in direct OA healthcare costs of 2.2% per year between 2015 and 2030, based on the changing age and gender distribution of the population (ie, excluding the effect of changes in population obesity). For comparison, we projected the same 2.2% annual increase between 2013 and 2038 when holding obesity rates constant at the 2013 level. A study in Canada projected increases in costs of OA treatment of 4.7% per year between 2010 and 2031;28 these higher rates are at least in part explained by differing assumptions about the course of future healthcare prices.

There is more literature on projected provision of TKR, although previous studies have reported widely-varying results due to differences in context, modelling assumptions and statistical methodology. Our projections (2.3% per year, or 1.9% excluding the effects of increasing obesity, which have often not been captured in past studies) are at the low end of published projections for the US (between 1.9% and 8.5%14,29), similar to those for the UK (1.6% to 2.8%30) and the Netherlands (1.7% to 2.7%32), and higher than those for Sweden (0.9%8). In the local context, our results are consistent with those of Hooper et al,13 when excluding the effect of increasing obesity rates, providing external validation of our model, and demonstrate the additional burden being placed on the health system by the continuing obesity epidemic.

The study has limitations relating to model structure and availability of data. The estimated costs associated with knee OA are not stratified by disease severity (other than the cost of TKR for end-stage OA), due to a lack of available data on the relationship between disease progression and treatment costs. As obesity is associated with both incidence and progression of knee OA, this limitation may have resulted in our projections underestimating the effects of increasing obesity on healthcare costs of OA. We have also modelled direct healthcare costs only, excluding non-health costs such as time off work or reduced productivity, informal care outside the health system, or equipment and aids to assist with daily living. These other costs may be substantial in OA; for example, a recent report on the costs of (all types of) arthritis in New Zealand found that productivity losses associated with arthritis were 125% of direct healthcare costs, and other non-health financial costs a further 202% of healthcare costs.5 These suggest that the healthcare costs reported here may represent only one-quarter of the total societal cost of knee OA. Lastly, the NZ-MOA model captures the most severely OA-affected knee only for each individual; the projections of TKR provision therefore relate only to the first TKR surgery per patient. According to our data from the New Zealand Joint Registry, these account for approximately 75% of all TKR in New Zealand.

Strengths of the study include the use of a validated, state-of-the-art computer simulation model, populated with comprehensive and reliable national-level data on disease prevalence, risk factors and TKR provision.18 By modelling the underlying drivers of future TKR incidence—initial prevalence of OA across the population, and ongoing disease incidence, structural progression and symptom severity—this approach may provide more reliable projections than statistical models relying on linear or exponential extrapolation of observed trends in TKR incidence rates far beyond the period of observed data. The sensitivity of data-driven statistical projections of TKR incidence to modelling assumptions about the ongoing trend is apparent in the wide variability in previously published estimates. Our projected increases in TKR incidence were generally higher than previously published estimates assuming constant incidence rates (within demographic strata), but (often substantially) lower than those assuming the continuation of observed short-run increasing trends.

The consistent results found in our cross-model validation exercise demonstrate the validity of the model for predicting future outcomes based on current practice patterns and risk factor distributions and reliable estimates from Statistics New Zealand of projected demographic change. The impact of increasing population obesity, additional to these validated projections based on demographic change, was derived from published estimates, for the New Zealand population, of future trends in population BMI, and international evidence on the relationship between BMI and OA incidence and progression. By combining these reliable sources of data, using our well-validated computer simulation model, we have been able to provide well-grounded, coherent and reliable projections of the increasing healthcare burden of knee OA in the New Zealand population.

Conclusion

The healthcare burden of knee OA in New Zealand will continue to grow over the next 25 years due to population ageing and increasing rates of obesity. Without changes in the provision of effective and cost-effective care throughout the disease course, the annual direct healthcare costs of knee OA will increase by 85% to $370 million by 2038, and an additional 4,000 TKR surgeries per year will be required.

Appendix A

This appendix describes the sources and derivations of the NZ-MOA model input parameters used in this study. For further details refer to Wilson & Abbott (2018)33 and the NZ-MOA Technical Manual, version 1.4.0 (available on request from the authors).

Demographic characteristics

Baseline age, gender and ethnicity are drawn from the discrete joint probability distribution of the New Zealand population, obtained from the 2013 New Zealand Census.34 The model uses an annual cycle, so age is incremented by one year each period; gender and ethnicity are assumed to be fixed throughout the lifetime of each individual. Mortality rates, stratified by age, gender and ethnicity, were obtained from the New Zealand Period Life Tables 2012–14,35 and further adjusted for the relative risk of mortality associated with obesity using data from the US population.36 Mortality rates were assumed to decrease by 1.75% per year for non-Māori and 2.25% per year for Māori until 2026 and remain constant thereafter.37

Body mass index

Baseline BMI is drawn from an age-, gender- and ethnicity-specific log-normal distribution (Table A1). Future trajectories of BMI are assumed to follow the population-average age trajectory of BMI, plus an annual trend increase reflecting increases in population obesity.38 The ongoing trend increase was drawn, in each loop of the PSA process, from a normal distribution with gender- and ethnicity-specific mean and variance (Table A2).

Table A1: Mean of population baseline BMI distribution, by age, gender and ethnicity.

Table A2: Trend increase in population BMI distribution.

Structural (radiographic) knee osteoarthritis

Baseline radiographic knee OA status is drawn from the age-, gender-, ethnicity- and obesity-specific prevalence of self-reported doctor-diagnosed all-site OA, derived from the New Zealand Health Survey 2013/14 (Table A3), and further adjusted for the ratio of knee OA to all-site OA39 and the ratio of radiographically-defined knee OA to self-reported diagnosis.40 Incidence was derived to be consistent with baseline prevalence rates, assuming no case fatality or remission in radiographic knee OA.

Table A3: Baseline prevalence of self-reported doctor-diagnosed all-site OA.


Values are coefficients from a logistic regression model (ie, log(Odds(OA))). Further adjustments to estimate radiographically-defined knee OA prevalence are: ratio of knee to all-site OA (men = 0.713, women = 0.665), odds ratio of radiographically-defined knee OA to self-reported diagnosis (1.30).

Annual progression of radiographic knee OA (defined by K-L grade 2/3/4) was obtained from US- and UK-based prospective cohort studies,41–43 stratified by gender, obesity and prior K-L grade (Table A4).

Health-related quality of life impacts of radiographic knee osteoarthritis

The HRQoL impacts of knee OA are modelled on each of the six dimensions of the SF-6D, which are then valued using the SF-6D preference-based utility scores.44 The average health utility loss associated with knee OA is 0.02 at K-L grade 2, 0.05 at K-L grade 3, and 0.10 at grade 4.

OA healthcare costs

The direct healthcare costs of OA treatment were based on the assumed provision of usual medical care as practiced in New Zealand (consisting of GP consultations, analgesic medication, and referrals to physical therapy for some patients),45 valued with New Zealand-specific reference prices sourced from PHARMAC’s Cost Resource Manual for HTA in New Zealand (Table A5).46

Table A5: Annual costs of OA-related healthcare.

Appendix B

Table B1: Projected healthcare costs and TKR provision, by year.

All values refer to the New Zealand adult population aged 25–99.

Table B2: Projected healthcare costs and TKR provision, by year, sex and ethnicity.

All values refer to the New Zealand adult population aged 25–99.

Figure B1: Observed and projected demand for total knee replacement in New Zealand, 1999–2038, cross-model comparison.

c

Source: NZ-MOA simulation model; NZJR 2016;22 Hooper et al 2014.13
Points show observed provision of TKR 1999–2013; dashed line the published projections from Hooper et al 2001–2026; and solid lines the new model projections 2013–2038, under the base case scenario (ongoing increase in population BMI based on current trends) and with no further increase in population BMI (ie, maintaining 2013 BMI distribution).
Inclusion criteria differ between data sources, resulting in the discontinuity between observed and modelled TKR provision in 2013. See the main manuscript for further discussion of modelling assumptions and inclusion criteria.
BMI indicates Body Mass Index, in kg/m2; NZJR, New Zealand Joint Registry; TKR, Total knee replacement surgery.  

Summary

Abstract

Aim

To estimate the healthcare costs and demand for total knee replacement (TKR) associated with knee osteoarthritis in New Zealand over the period 2013-2038 and the contribution of increasing obesity rates to these costs.

Method

We used the NZ-MOA computer simulation model of knee osteoarthritis in the New Zealand population. Osteoarthritis-related healthcare costs and TKR incidence were modelled for a 25-year horizon, for a starting cohort drawn from the 2013 New Zealand population. Population obesity projections were used to estimate the life-course of cohort obesity. Per-person projected outcomes were multiplied by national demographic population projections to obtain total population projections.

Results

Healthcare costs of knee osteoarthritis were projected to increase from NZ$199 million in 2013 to $370 million in 2038. Annual TKR incidence was projected to increase from 5,070 to 9,040 over the same period. Projected increases in population obesity rates (above the obesity prevalence seen in 2013) accounted for 25% and 47% of the projected increase in per-capita healthcare costs and TKR provision rates, respectively.

Conclusion

The healthcare burden of knee OA will continue to increase for the foreseeable future. Public health measures to reduce further increases in population obesity rates would contribute to slowing this rising burden.

Author Information

Ross Wilson, Research Fellow, Centre for Musculoskeletal Outcomes Research, University of Otago, Dunedin; J Haxby Abbott, Research Professor, Centre for Musculoskeletal Outcomes Research, University of Otago, Dunedin.

Acknowledgements

Correspondence

Dr Ross Wilson, Centre for Musculoskeletal Outcomes Research, Department of Surgical Sciences, University of Otago, PO Box 56, Dunedin 9054.

Correspondence Email

ross.wilson@otago.ac.nz

Competing Interests

Dr Wilson reports grants from Health Research Council of New Zealand during the conduct of the study.

  1. Cross M, Smith E, Hoy D, et al. The global burden of hip and knee osteoarthritis: Estimates from the Global Burden of Disease 2010 study. Ann Rheum Dis. 2014; 73:1323–30.
  2. Ministry of Health, Health loss in New Zealand 1990-2013: A report from the New Zealand Burden of Diseases, Injuries and Risk Factor Study. Wellington, New Zealand: Ministry of Health; 2016.
  3. GBD 2015 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990-2015: A systematic analysis for the Global Burden of Disease Study 2015. Lancet. 2016; 388:1545–602.
  4. Institute for Health Metrics and Evaluation. GBD Compare. Available at: http://gbd2016.healthdata.org/gbd-compare/ (accessed 2019/04/17); 2019.
  5. Deloitte Access Economics. The economic cost of arthritis in New Zealand in 2018. Arthritis New Zealand; 2018 Aug.
  6. Bruyère O, Ethgen O, Neuprez A, et al. Health-related quality of life after total knee or hip replacement for osteoarthritis: A 7-year prospective study. Arch Orthop Trauma Surg. 2012; 132:1583–7.
  7. Ethgen O, Bruyère O, Richy F, et al. Health-related quality of life in total hip and total knee arthroplasty: A qualitative and systematic review of the literature. J Bone Joint Surg. 2004; 86-A:963–74.
  8. Robertsson O, Dunbar MJ, Knutson K, Lidgren L. Past incidence and future demand for knee arthroplasty in Sweden: A report from the Swedish Knee Arthroplasty Register regarding the effect of past and future population changes on the number of arthroplasties performed. Acta Orthop Scand. 2000; 71:376–80.
  9. Kim H-A, Kim S, Seo YI, et al. The epidemiology of total knee replacement in South Korea: National registry data. Rheumatology. 2008; 47:88–91.
  10. Pedersen AB, Johnsen SP, Overgaard S, et al. Total hip arthroplasty in Denmark: Incidence of primary operations and revisions during 1996–2002 and estimated future demands. Acta Orthop. 2005; 76:182–9.
  11. Ravi B, Croxford R, Reichmann WM, et al. The changing demographics of total joint arthroplasty recipients in the United States and Ontario from 2001 to 2007. Best Pract Res Clin Rheumatol. 2012; 26:637–47.
  12. Kurtz SM, Lau E, Ong K, et al. Future young patient demand for primary and revision joint replacement: National projections from 2010 to 2030. Clin Orthop Relat Res. 2009; 467:2606–12.
  13. Hooper G, Lee AJ-J, Rothwell A, Frampton C. Current trends and projections in the utilisation rates of hip and knee replacement in New Zealand from 2001 to 2026. NZ Med J. 2014; 127:82–93.
  14. Kurtz S, Ong K, Lau E, et al. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am. 2007; 89-A:780–5.
  15. Nemes S, Gordon M, Rogmark C, Rolfson O. Projections of total hip replacement in Sweden from 2013 to 2030. Acta Orthop. 2014; 85:238–43.
  16. Statistics New Zealand. National population projections: 2016(Base)–2068. Available at: http://www.stats.govt.nz/information-releases/national-population-projections-2016base2068 (accessed 2019/04/04); 2016.
  17. Wilson R, Abbott JH. Age, period and cohort effects on body mass index in New Zealand, 1997–2038. Aust NZ J Public Health. 2018; 42:396–402.
  18. Wilson R, Abbott JH. Development and validation of a new population-based simulation model of osteoarthritis in New Zealand. Osteoarthritis Cartilage. 2018; 26:531–9.
  19. Statistics New Zealand. National ethnic population projections: 2013(Base)–2038 (update). Available at: http://www.stats.govt.nz/information-releases/national-ethnic-population-projections-2013base2038-update (accessed 2018/11/22); 2017.
  20. Pinto D, Robertson MC, Abbott JH, et al. Manual therapy, exercise therapy, or both, in addition to usual care, for osteoarthritis of the hip or knee. 2: Economic evaluation alongside a randomized controlled trial. Osteoarthritis Cartilage. 2013; 21:1504–13.
  21. PHARMAC, Cost Resource Manual version 2.2. PHARMAC; 2015.
  22. New Zealand Joint Registry, Sixteen year report: January 1999 to December 2015. New Zealand Joint Registry; 2016 Oct.
  23. Gwynne-Jones D. Quantifying the demand for hip and knee replacement in Otago, New Zealand. NZ Med J. 2013; 126:7–17.
  24. Harcombe H, Davie G, Derrett S, et al. Equity of publicly-funded hip and knee joint replacement surgery in New Zealand: Results from a national observational study. NZ Med J. 2016; 129:8–18.
  25. Abbott JH, Wilson R, Pinto D, et al. Incremental clinical effectiveness and cost effectiveness of providing supervised physiotherapy in addition to usual medical care in patients with osteoarthritis of the hip or knee: 2-year results of the MOA randomised controlled trial. Osteoarthritis Cartilage. 2019; 27:424–34.
  26. Teoh LSG, Eyles JP, Makovey J, et al. Observational study of the impact of an individualized multidisciplinary chronic care program for hip and knee osteoarthritis treatment on willingness for surgery. Int J Rheum Dis. 2017; 20:1383–92.
  27. Ackerman IN, Pratt C, Gorelik A, Liew D. Projected burden of osteoarthritis and rheumatoid arthritis in Australia: A population-level analysis. Arthritis Care Res. 2018; 70:877–83.
  28. Sharif B, Kopec J, Bansback N, et al. Projecting the direct cost burden of osteoarthritis in Canada using a microsimulation model. Osteoarthritis Cartilage. 2015; 23:1654–63.
  29. Inacio MCS, Paxton EW, Graves SE, et al. Projected increase in total knee arthroplasty in the United States – an alternative projection model. Osteoarthritis Cartilage. 2017; 25:1797–803.
  30. Culliford D, Maskell J, Judge A, et al. Future projections of total hip and knee arthroplasty in the UK: Results from the UK Clinical Practice Research Datalink. Osteoarthritis Cartilage. 2015; 23:594–600.
  31. Patel A, Pavlou G, Mújica-Mota RE, Toms AD. The epidemiology of revision total knee and hip arthroplasty in England and Wales: A comparative analysis with projections for the United States. A study using the National Joint Registry dataset. Bone Joint J. 2015; 97-B:1076–81.
  32. Otten R, Roermund PM van, Picavet HS. Trends in the number of knee and hip arthoplasties: Considerably more knee and hip prostheses due to osteoarthritis in 2030. Ned Tijdschr Geneeskd. 2010; 154:A1534.
  33. Wilson R, Abbott JH. Development and validation of a new population-based simulation model of osteoarthritis in New Zealand. Osteoarthritis Cartilage. 2018; 26:531–9.
  34. Statistics New Zealand. 2013 Census. Wellington, New Zealand: Statistics New Zealand; 2014.
  35. Statistics New Zealand. New Zealand period life tables, 2012–2014. Wellington, New Zealand: Statistics New Zealand; 2015.
  36. Berrington de Ganzalez A, Hartge P, Cerhan JR, et al. Body mass index and mortality among 1.46 million white adults. N Engl J Med. 2010; 363(23):2211–9.
  37. Blakely T, Foster R, Wilson N. Burden of Disease Epidemiology, Equity, and Cost-Effectiveness (BODE3) study protocol, version 2.1. Wellington, New Zealand: University of Otago; 2012. (Public health monograph series; vol. 30).
  38. Wilson R, Abbott JH. Age, period and cohort effects on body mass index in New Zealand, 1997–2038. Aust NZ J Public Health. 2018; 42:396–402.
  39. Abbott JH, Usiskin IM, Wilson R, et al. The quality-of-life burden of knee osteoarthritis in New Zealand adults: A model-based evaluation. PLOS ONE. 2017; 12(10):e0185676.
  40. GBD 2015 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990-2015: A systematic analysis for the Global Burde`n of Disease Study 2015. Lancet. 2016; 388:1545–602.
  41. Holt HL, Katz JN, Reichmann WM, et al. Forecasting the burden of advanced knee osteoarthritis over a 10-year period in a cohort of 60–64 year-old US adults. Osteoarthritis Cartilage. 2011; 19:44–50.
  42. Jordan JM, Helmick CG, Renner JB, et al. Prevalence of knee symptoms and radiographic and symptomatic knee osteoarthritis in African Americans and Caucasians: The Johnston County Osteoarthritis Project. J Rheumatol. 2007; 34:172–80.
  43. Cooper C, Snow S, McAlindon TE, et al. Risk factors for the incidence and progression of radiographic knee osteoarthritis. Arthritis Rheum. 2000; 43(5):995–1000.
  44. Brazier JE, Roberts J. The estimation of a preference-based measure of health from the SF-12. Med Care. 2004; 42(9):851–9.
  45. Pinto D, Robertson MC, Abbott JH, et al. Manual therapy, exercise therapy, or both, in addition to usual care, for osteoarthritis of the hip or knee. 2: Economic evaluation alongside a randomised controlled trial. Osteoarthritis Cartilage. 2013; 21:1504–13.
  46. Pharmaceutical Management Agency (PHARMAC). Cost Resource Manual, version 2.2. Wellington, New Zealand: PHARMAC; 2015.

For the PDF of this article,
contact nzmj@nzma.org.nz

View Article PDF

Osteoarthritis (OA) is a common and debilitating chronic disease and one of the leading causes of disability in New Zealand and worldwide.1–3 Knee OA is the most common form of OA, affecting as much as one-third of the population at some point in their lifetime. Globally, OA is responsible for more than 16 million disability-adjusted life years (DALYs) and is the 12th leading cause of worldwide disability.4 In New Zealand, osteoarthritis accounted for more than 15,000 DALYs in 2018.5 In addition to the health losses associated with OA, it also accounts for a substantial economic burden on the health system and the wider economy. In New Zealand, the total annual financial cost of arthritis (of all types) is estimated to be $4.2 billion, including health system costs of $990 million; productivity costs of $1.2 billion, through reduced employment, time off work and impaired performance at work; and other costs, such as non-health sector care and aids and modifications to support independent living for people with arthritis, of $2 billion.5

Total knee replacement (TKR) surgery is a common and successful operation to reduce pain and improve HRQoL in patients with advanced knee OA.6,7 Provision of TKR has been increasing across the world in recent decades.8–11 However, TKR is a costly procedure and high provision rates place strain on limited public healthcare resources. Furthermore, access to TKR, particularly in the public healthcare system, is limited by capacity constraints due to the availability of a suitably-trained workforce and surgical facilities.

The two most important risk factors for incidence and progression of knee OA are age and obesity. In New Zealand, as elsewhere, the combination of an ageing population and increasing rates of obesity is therefore expected to result in continuing increases in demand for OA healthcare.8,10,12–15 The population aged over 65 is expected grow by 40 percent over the next decade,16 while obesity rates are projected to increase to close to 50% of the adult population by the late 2030s.17 Understanding the implications of these population changes is critical for successful health system and workforce planning.

The aims of this study are (1) to estimate the projected healthcare costs of osteoarthritis care in New Zealand for the period 2013–2038, (2) to estimate the demand for TKR surgery in New Zealand over the same period, and (3) to assess the contribution of projected increases in population obesity to future healthcare expenditure and TKR demand.

Methods

We used the New Zealand Management of Osteoarthritis (NZ-MOA) model, a state-transition microsimulation model of the incidence, progression, health impact and healthcare costs of knee OA in New Zealand.18 The model generates a hypothetical cohort drawn from the 2013 New Zealand population distribution; runs each simulated individual through a sequence of annual health state transitions, defined by obesity, knee OA status (Kellgren-Lawrence grade), health-related quality of life (HRQoL) outcomes and treatment pathway; and computes the resulting per-capita healthcare costs, quality-adjusted life years (QALYs) and treatment utilisation. By running the same cohort through multiple scenarios, the incremental change in outcomes attributable to knee OA prevalence, change in risk factors or change in treatment patterns can be estimated.

We computed model outcomes under two scenarios: continuation of projected trends in population obesity levels17 and holding population obesity constant at 2013 levels. We generated simulation cohorts of 10,000 individuals for each age- (in five-year age groups), gender- and ethnicity-specific population subgroup. For each subgroup, we ran each scenario for a modelled 25-year time horizon (2013–2038) and recorded the age-, gender- and ethnicity-specific OA-related healthcare expenditure and TKR incidence in each year. Projections of healthcare costs and joint replacement provision at the population level were calculated by multiplying the age-, gender- and ethnicity-specific per-person projected outcomes from the simulation model by Statistics New Zealand’s published national population projections for each future period.19 This whole process was repeated 1,000 times, each with a new random draw from the distribution of obesity projections, resulting in a total simulated population of 760 million individuals to provide stable and precise estimates of the outcomes. We calculated the mean healthcare costs and joint replacement incidence in each year across the 1,000 cycles to provide the projected point estimates for each outcome, and the 2.5th and 97.5th percentiles to construct 95% uncertainty intervals for the contribution of increasing obesity levels (ie, undertaking probabilistic sensitivity analysis).

Baseline population model input parameters and sources have been described previously,18 and are summarised in Appendix A. Costs of treatment were based on provision of usual publicly and privately provided medical care as practiced in New Zealand,20 valued with New Zealand-specific reference prices21 (see Appendix Table A5). TKR incidence rates in the New Zealand population were obtained from the New Zealand Joint Registry,22 which covers almost all privately and publicly funded joint replacement surgeries in New Zealand, and used to calibrate the baseline modelled TKR provision rates. Projected increases in population obesity rates, stratified by age, gender and ethnicity, were obtained from a previously published age-period-cohort model of body mass index (BMI) in the New Zealand population.17

Cross-model validation was conducted by comparing our projected increases in the provision of TKR with those previously published in the Journal by Gary Hooper and colleagues (2014).13 Those estimates were based on different inclusion criteria than our model: all TKR were included for any primary diagnosis, whereas our model includes only one OA knee per person (ie, only the first primary TKR performed), and includes only TKR performed for a primary diagnosis of knee OA. The total number of TKR performed annually should therefore be lower in our model projections; to allow direct cross-model comparison, we normalised all estimates to an index value of 100 in 2011 (the base year in Hooper et al’s projections), to compare the total percentage increase in TKR provision between models. (The observed increase in TKR provision between 2011 and 2013 was used to determine the baseline (2013) index value for the NZ-MOA model, as the model was calibrated to observed 2013 data.)

Results

In 2013 (the base year for the NZ-MOA model), the total direct healthcare costs of knee OA were estimated to be NZ$199 million (1NZD≈0.82USD), and 5,070 patients had a first total joint replacement for knee OA. Women accounted for $110 million (55%) of the total costs and 2,460 TKR (48%); Māori for $18 million (9%) of the total costs and 490 TKR (10%).

The healthcare costs associated with OA treatment were projected to increase to $370 million per year in 2038 (at constant 2013 prices; Figure 1), an increase of 85%. Per-capita treatment costs (for the adult population aged 25 and over) were projected to increase from $69 to $90. TKR incidence was projected to increase to 9,040 per year in 2038 (Figure 2). The provision rate was projected to increase from 174 to 221 per 100,000 population per year.

Figure 1: Projected healthcare costs of knee osteoarthritis in New Zealand, 2013–2038.

c

Source: NZ-MOA simulation model. Lines show the projected increase in annual healthcare expenditure on OA treatment under the base case scenario (ongoing increase in population BMI based on current trends) and with no further increase in population BMI (ie, maintaining 2013 BMI distribution), relative to 2013 baseline level. ‘Attributable to BMI increase’ is the difference between the two lines, representing the incremental OA-related healthcare expenditure attributable to continued increases in population BMI relative to the 2013 level. The shaded uncertainty interval represents the 95% confidence interval for projected annual increase in population BMI. Right-hand axis shows total projected annual healthcare expenditure in each period for the base case and no BMI increase scenarios.
BMI indicates Body Mass Index; OA, Knee osteoarthritis.

Figure 2: Projected demand for first total knee replacement for knee osteoarthritis in New Zealand, 2013–2038.

c

Source: NZ-MOA simulation model. Lines show the projected increase in annual provision of TKR under the base case scenario (ongoing increase in population BMI based on current trends) and with no further increase in population BMI (ie, maintaining 2013 BMI distribution), relative to 2013 baseline level. ‘Attributable to BMI increase’ is the difference between the two lines, representing the incremental TKR provision attributable to continued increases in population BMI relative to the 2013 level. The shaded uncertainty interval represents the 95% confidence interval for projected annual increase in population BMI. Right-hand axis shows total projected annual provision in each period for the base case and no BMI increase scenarios.
BMI indicates Body Mass Index; TKR, Total knee replacement surgery.

In 2038, women were projected to account for $210 million (57%) of the total healthcare costs and 4,760 TKR (53%); Māori for $44 million (12%) of the total costs and 1,120 TKR (12%). Full results for projected costs and TKR incidence for each year 2013–2038, including ethnic and gender breakdowns, are reported in Tables B1 and B2 in the Appendix.

Further increases in population BMI, relative to the baseline distribution in 2013, accounted for 25% of the projected increase in per-capita OA-related healthcare costs and 47% of the increase in TKR provision rates between 2013 and 2038. By 2038, this accounts for an additional $22 million and 880 TKR per year. Over the 25-year time horizon, projected increases in BMI, relative to baseline levels, will result in an additional $231 million in OA healthcare costs and 9,880 TKR required.

The increase in TKR provision projected by our model, in the absence of further increases in population obesity, was very close to that previously reported, using a different modelling approach, by Hooper et al (2014) (Figure 3).13 Between 2011 and 2026, TKR provision was projected to increase by 48.8%, compared to 49.0% reported in the previously-published estimates. Allowing for projected increases in population obesity, this increased to 56.6% in our model, and 90.8% by 2038. The total numbers of projected TKR (without normalisation) from both models are shown in Figure B1 in the Appendix.

Figure 3: Observed and projected demand for total knee replacement in New Zealand, 1999–2038, cross-model comparison.

c

Source: NZ-MOA simulation model; NZJR 2016;22 Hooper et al 2014.13 Points show observed provision of TKR 1999–2013; dashed line the published projections from Hooper et al 2011–2026; and solid lines the new model projections 2013–2038, under the base case scenario (ongoing increase in population BMI based on current trends) and with no further increase in population BMI (ie, maintaining 2013 BMI distribution). The absolute numbers in each series differ due to different inclusion criteria: all values are normalised to 2011 = 100 for consistent comparisons.
BMI indicates Body Mass Index, in kg/m2; NZJR, New Zealand Joint Registry; TKR, Total knee replacement surgery.

Discussion

Treatment of knee OA cost the New Zealand healthcare system $200 million in 2013, an amount projected to increase to $370 million over 25 years. Demand for TKR is projected to increase by almost 80% over the same period, requiring an additional 4,000 surgeries to be performed annually by 2038. After adjusting for population growth, projected increases in obesity rates account for one-quarter of the increase in per-capita costs and almost half of the increase in TKR provision rates over 25 years.

These results highlight the need for effective—and cost-effective—treatment of OA throughout the disease course. Our projections were calculated under the assumption that patients’ access to TKR (and other treatments) remains at the same level as in 2013. Access to TKR is already rationed through the public health system, with many local health systems unable to offer surgery to all patients who would benefit;23,24 improving access to TKR would require further resourcing beyond that suggested by our projections. Effective, low-cost, early interventions, such as exercise therapy, can alleviate symptoms, improve patients’ quality of life and reduce the need for costly treatment, such as TKR, later in the disease course.25,26 Improving access to such treatments may help to mitigate the increasing burden on the health system associated with rising rates of OA.

There is limited existing research on the future healthcare costs of OA. Our projected increases in costs are very similar to those reported recently for Australia,27 which had a projected increase in direct OA healthcare costs of 2.2% per year between 2015 and 2030, based on the changing age and gender distribution of the population (ie, excluding the effect of changes in population obesity). For comparison, we projected the same 2.2% annual increase between 2013 and 2038 when holding obesity rates constant at the 2013 level. A study in Canada projected increases in costs of OA treatment of 4.7% per year between 2010 and 2031;28 these higher rates are at least in part explained by differing assumptions about the course of future healthcare prices.

There is more literature on projected provision of TKR, although previous studies have reported widely-varying results due to differences in context, modelling assumptions and statistical methodology. Our projections (2.3% per year, or 1.9% excluding the effects of increasing obesity, which have often not been captured in past studies) are at the low end of published projections for the US (between 1.9% and 8.5%14,29), similar to those for the UK (1.6% to 2.8%30) and the Netherlands (1.7% to 2.7%32), and higher than those for Sweden (0.9%8). In the local context, our results are consistent with those of Hooper et al,13 when excluding the effect of increasing obesity rates, providing external validation of our model, and demonstrate the additional burden being placed on the health system by the continuing obesity epidemic.

The study has limitations relating to model structure and availability of data. The estimated costs associated with knee OA are not stratified by disease severity (other than the cost of TKR for end-stage OA), due to a lack of available data on the relationship between disease progression and treatment costs. As obesity is associated with both incidence and progression of knee OA, this limitation may have resulted in our projections underestimating the effects of increasing obesity on healthcare costs of OA. We have also modelled direct healthcare costs only, excluding non-health costs such as time off work or reduced productivity, informal care outside the health system, or equipment and aids to assist with daily living. These other costs may be substantial in OA; for example, a recent report on the costs of (all types of) arthritis in New Zealand found that productivity losses associated with arthritis were 125% of direct healthcare costs, and other non-health financial costs a further 202% of healthcare costs.5 These suggest that the healthcare costs reported here may represent only one-quarter of the total societal cost of knee OA. Lastly, the NZ-MOA model captures the most severely OA-affected knee only for each individual; the projections of TKR provision therefore relate only to the first TKR surgery per patient. According to our data from the New Zealand Joint Registry, these account for approximately 75% of all TKR in New Zealand.

Strengths of the study include the use of a validated, state-of-the-art computer simulation model, populated with comprehensive and reliable national-level data on disease prevalence, risk factors and TKR provision.18 By modelling the underlying drivers of future TKR incidence—initial prevalence of OA across the population, and ongoing disease incidence, structural progression and symptom severity—this approach may provide more reliable projections than statistical models relying on linear or exponential extrapolation of observed trends in TKR incidence rates far beyond the period of observed data. The sensitivity of data-driven statistical projections of TKR incidence to modelling assumptions about the ongoing trend is apparent in the wide variability in previously published estimates. Our projected increases in TKR incidence were generally higher than previously published estimates assuming constant incidence rates (within demographic strata), but (often substantially) lower than those assuming the continuation of observed short-run increasing trends.

The consistent results found in our cross-model validation exercise demonstrate the validity of the model for predicting future outcomes based on current practice patterns and risk factor distributions and reliable estimates from Statistics New Zealand of projected demographic change. The impact of increasing population obesity, additional to these validated projections based on demographic change, was derived from published estimates, for the New Zealand population, of future trends in population BMI, and international evidence on the relationship between BMI and OA incidence and progression. By combining these reliable sources of data, using our well-validated computer simulation model, we have been able to provide well-grounded, coherent and reliable projections of the increasing healthcare burden of knee OA in the New Zealand population.

Conclusion

The healthcare burden of knee OA in New Zealand will continue to grow over the next 25 years due to population ageing and increasing rates of obesity. Without changes in the provision of effective and cost-effective care throughout the disease course, the annual direct healthcare costs of knee OA will increase by 85% to $370 million by 2038, and an additional 4,000 TKR surgeries per year will be required.

Appendix A

This appendix describes the sources and derivations of the NZ-MOA model input parameters used in this study. For further details refer to Wilson & Abbott (2018)33 and the NZ-MOA Technical Manual, version 1.4.0 (available on request from the authors).

Demographic characteristics

Baseline age, gender and ethnicity are drawn from the discrete joint probability distribution of the New Zealand population, obtained from the 2013 New Zealand Census.34 The model uses an annual cycle, so age is incremented by one year each period; gender and ethnicity are assumed to be fixed throughout the lifetime of each individual. Mortality rates, stratified by age, gender and ethnicity, were obtained from the New Zealand Period Life Tables 2012–14,35 and further adjusted for the relative risk of mortality associated with obesity using data from the US population.36 Mortality rates were assumed to decrease by 1.75% per year for non-Māori and 2.25% per year for Māori until 2026 and remain constant thereafter.37

Body mass index

Baseline BMI is drawn from an age-, gender- and ethnicity-specific log-normal distribution (Table A1). Future trajectories of BMI are assumed to follow the population-average age trajectory of BMI, plus an annual trend increase reflecting increases in population obesity.38 The ongoing trend increase was drawn, in each loop of the PSA process, from a normal distribution with gender- and ethnicity-specific mean and variance (Table A2).

Table A1: Mean of population baseline BMI distribution, by age, gender and ethnicity.

Table A2: Trend increase in population BMI distribution.

Structural (radiographic) knee osteoarthritis

Baseline radiographic knee OA status is drawn from the age-, gender-, ethnicity- and obesity-specific prevalence of self-reported doctor-diagnosed all-site OA, derived from the New Zealand Health Survey 2013/14 (Table A3), and further adjusted for the ratio of knee OA to all-site OA39 and the ratio of radiographically-defined knee OA to self-reported diagnosis.40 Incidence was derived to be consistent with baseline prevalence rates, assuming no case fatality or remission in radiographic knee OA.

Table A3: Baseline prevalence of self-reported doctor-diagnosed all-site OA.


Values are coefficients from a logistic regression model (ie, log(Odds(OA))). Further adjustments to estimate radiographically-defined knee OA prevalence are: ratio of knee to all-site OA (men = 0.713, women = 0.665), odds ratio of radiographically-defined knee OA to self-reported diagnosis (1.30).

Annual progression of radiographic knee OA (defined by K-L grade 2/3/4) was obtained from US- and UK-based prospective cohort studies,41–43 stratified by gender, obesity and prior K-L grade (Table A4).

Health-related quality of life impacts of radiographic knee osteoarthritis

The HRQoL impacts of knee OA are modelled on each of the six dimensions of the SF-6D, which are then valued using the SF-6D preference-based utility scores.44 The average health utility loss associated with knee OA is 0.02 at K-L grade 2, 0.05 at K-L grade 3, and 0.10 at grade 4.

OA healthcare costs

The direct healthcare costs of OA treatment were based on the assumed provision of usual medical care as practiced in New Zealand (consisting of GP consultations, analgesic medication, and referrals to physical therapy for some patients),45 valued with New Zealand-specific reference prices sourced from PHARMAC’s Cost Resource Manual for HTA in New Zealand (Table A5).46

Table A5: Annual costs of OA-related healthcare.

Appendix B

Table B1: Projected healthcare costs and TKR provision, by year.

All values refer to the New Zealand adult population aged 25–99.

Table B2: Projected healthcare costs and TKR provision, by year, sex and ethnicity.

All values refer to the New Zealand adult population aged 25–99.

Figure B1: Observed and projected demand for total knee replacement in New Zealand, 1999–2038, cross-model comparison.

c

Source: NZ-MOA simulation model; NZJR 2016;22 Hooper et al 2014.13
Points show observed provision of TKR 1999–2013; dashed line the published projections from Hooper et al 2001–2026; and solid lines the new model projections 2013–2038, under the base case scenario (ongoing increase in population BMI based on current trends) and with no further increase in population BMI (ie, maintaining 2013 BMI distribution).
Inclusion criteria differ between data sources, resulting in the discontinuity between observed and modelled TKR provision in 2013. See the main manuscript for further discussion of modelling assumptions and inclusion criteria.
BMI indicates Body Mass Index, in kg/m2; NZJR, New Zealand Joint Registry; TKR, Total knee replacement surgery.  

Summary

Abstract

Aim

To estimate the healthcare costs and demand for total knee replacement (TKR) associated with knee osteoarthritis in New Zealand over the period 2013-2038 and the contribution of increasing obesity rates to these costs.

Method

We used the NZ-MOA computer simulation model of knee osteoarthritis in the New Zealand population. Osteoarthritis-related healthcare costs and TKR incidence were modelled for a 25-year horizon, for a starting cohort drawn from the 2013 New Zealand population. Population obesity projections were used to estimate the life-course of cohort obesity. Per-person projected outcomes were multiplied by national demographic population projections to obtain total population projections.

Results

Healthcare costs of knee osteoarthritis were projected to increase from NZ$199 million in 2013 to $370 million in 2038. Annual TKR incidence was projected to increase from 5,070 to 9,040 over the same period. Projected increases in population obesity rates (above the obesity prevalence seen in 2013) accounted for 25% and 47% of the projected increase in per-capita healthcare costs and TKR provision rates, respectively.

Conclusion

The healthcare burden of knee OA will continue to increase for the foreseeable future. Public health measures to reduce further increases in population obesity rates would contribute to slowing this rising burden.

Author Information

Ross Wilson, Research Fellow, Centre for Musculoskeletal Outcomes Research, University of Otago, Dunedin; J Haxby Abbott, Research Professor, Centre for Musculoskeletal Outcomes Research, University of Otago, Dunedin.

Acknowledgements

Correspondence

Dr Ross Wilson, Centre for Musculoskeletal Outcomes Research, Department of Surgical Sciences, University of Otago, PO Box 56, Dunedin 9054.

Correspondence Email

ross.wilson@otago.ac.nz

Competing Interests

Dr Wilson reports grants from Health Research Council of New Zealand during the conduct of the study.

  1. Cross M, Smith E, Hoy D, et al. The global burden of hip and knee osteoarthritis: Estimates from the Global Burden of Disease 2010 study. Ann Rheum Dis. 2014; 73:1323–30.
  2. Ministry of Health, Health loss in New Zealand 1990-2013: A report from the New Zealand Burden of Diseases, Injuries and Risk Factor Study. Wellington, New Zealand: Ministry of Health; 2016.
  3. GBD 2015 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990-2015: A systematic analysis for the Global Burden of Disease Study 2015. Lancet. 2016; 388:1545–602.
  4. Institute for Health Metrics and Evaluation. GBD Compare. Available at: http://gbd2016.healthdata.org/gbd-compare/ (accessed 2019/04/17); 2019.
  5. Deloitte Access Economics. The economic cost of arthritis in New Zealand in 2018. Arthritis New Zealand; 2018 Aug.
  6. Bruyère O, Ethgen O, Neuprez A, et al. Health-related quality of life after total knee or hip replacement for osteoarthritis: A 7-year prospective study. Arch Orthop Trauma Surg. 2012; 132:1583–7.
  7. Ethgen O, Bruyère O, Richy F, et al. Health-related quality of life in total hip and total knee arthroplasty: A qualitative and systematic review of the literature. J Bone Joint Surg. 2004; 86-A:963–74.
  8. Robertsson O, Dunbar MJ, Knutson K, Lidgren L. Past incidence and future demand for knee arthroplasty in Sweden: A report from the Swedish Knee Arthroplasty Register regarding the effect of past and future population changes on the number of arthroplasties performed. Acta Orthop Scand. 2000; 71:376–80.
  9. Kim H-A, Kim S, Seo YI, et al. The epidemiology of total knee replacement in South Korea: National registry data. Rheumatology. 2008; 47:88–91.
  10. Pedersen AB, Johnsen SP, Overgaard S, et al. Total hip arthroplasty in Denmark: Incidence of primary operations and revisions during 1996–2002 and estimated future demands. Acta Orthop. 2005; 76:182–9.
  11. Ravi B, Croxford R, Reichmann WM, et al. The changing demographics of total joint arthroplasty recipients in the United States and Ontario from 2001 to 2007. Best Pract Res Clin Rheumatol. 2012; 26:637–47.
  12. Kurtz SM, Lau E, Ong K, et al. Future young patient demand for primary and revision joint replacement: National projections from 2010 to 2030. Clin Orthop Relat Res. 2009; 467:2606–12.
  13. Hooper G, Lee AJ-J, Rothwell A, Frampton C. Current trends and projections in the utilisation rates of hip and knee replacement in New Zealand from 2001 to 2026. NZ Med J. 2014; 127:82–93.
  14. Kurtz S, Ong K, Lau E, et al. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am. 2007; 89-A:780–5.
  15. Nemes S, Gordon M, Rogmark C, Rolfson O. Projections of total hip replacement in Sweden from 2013 to 2030. Acta Orthop. 2014; 85:238–43.
  16. Statistics New Zealand. National population projections: 2016(Base)–2068. Available at: http://www.stats.govt.nz/information-releases/national-population-projections-2016base2068 (accessed 2019/04/04); 2016.
  17. Wilson R, Abbott JH. Age, period and cohort effects on body mass index in New Zealand, 1997–2038. Aust NZ J Public Health. 2018; 42:396–402.
  18. Wilson R, Abbott JH. Development and validation of a new population-based simulation model of osteoarthritis in New Zealand. Osteoarthritis Cartilage. 2018; 26:531–9.
  19. Statistics New Zealand. National ethnic population projections: 2013(Base)–2038 (update). Available at: http://www.stats.govt.nz/information-releases/national-ethnic-population-projections-2013base2038-update (accessed 2018/11/22); 2017.
  20. Pinto D, Robertson MC, Abbott JH, et al. Manual therapy, exercise therapy, or both, in addition to usual care, for osteoarthritis of the hip or knee. 2: Economic evaluation alongside a randomized controlled trial. Osteoarthritis Cartilage. 2013; 21:1504–13.
  21. PHARMAC, Cost Resource Manual version 2.2. PHARMAC; 2015.
  22. New Zealand Joint Registry, Sixteen year report: January 1999 to December 2015. New Zealand Joint Registry; 2016 Oct.
  23. Gwynne-Jones D. Quantifying the demand for hip and knee replacement in Otago, New Zealand. NZ Med J. 2013; 126:7–17.
  24. Harcombe H, Davie G, Derrett S, et al. Equity of publicly-funded hip and knee joint replacement surgery in New Zealand: Results from a national observational study. NZ Med J. 2016; 129:8–18.
  25. Abbott JH, Wilson R, Pinto D, et al. Incremental clinical effectiveness and cost effectiveness of providing supervised physiotherapy in addition to usual medical care in patients with osteoarthritis of the hip or knee: 2-year results of the MOA randomised controlled trial. Osteoarthritis Cartilage. 2019; 27:424–34.
  26. Teoh LSG, Eyles JP, Makovey J, et al. Observational study of the impact of an individualized multidisciplinary chronic care program for hip and knee osteoarthritis treatment on willingness for surgery. Int J Rheum Dis. 2017; 20:1383–92.
  27. Ackerman IN, Pratt C, Gorelik A, Liew D. Projected burden of osteoarthritis and rheumatoid arthritis in Australia: A population-level analysis. Arthritis Care Res. 2018; 70:877–83.
  28. Sharif B, Kopec J, Bansback N, et al. Projecting the direct cost burden of osteoarthritis in Canada using a microsimulation model. Osteoarthritis Cartilage. 2015; 23:1654–63.
  29. Inacio MCS, Paxton EW, Graves SE, et al. Projected increase in total knee arthroplasty in the United States – an alternative projection model. Osteoarthritis Cartilage. 2017; 25:1797–803.
  30. Culliford D, Maskell J, Judge A, et al. Future projections of total hip and knee arthroplasty in the UK: Results from the UK Clinical Practice Research Datalink. Osteoarthritis Cartilage. 2015; 23:594–600.
  31. Patel A, Pavlou G, Mújica-Mota RE, Toms AD. The epidemiology of revision total knee and hip arthroplasty in England and Wales: A comparative analysis with projections for the United States. A study using the National Joint Registry dataset. Bone Joint J. 2015; 97-B:1076–81.
  32. Otten R, Roermund PM van, Picavet HS. Trends in the number of knee and hip arthoplasties: Considerably more knee and hip prostheses due to osteoarthritis in 2030. Ned Tijdschr Geneeskd. 2010; 154:A1534.
  33. Wilson R, Abbott JH. Development and validation of a new population-based simulation model of osteoarthritis in New Zealand. Osteoarthritis Cartilage. 2018; 26:531–9.
  34. Statistics New Zealand. 2013 Census. Wellington, New Zealand: Statistics New Zealand; 2014.
  35. Statistics New Zealand. New Zealand period life tables, 2012–2014. Wellington, New Zealand: Statistics New Zealand; 2015.
  36. Berrington de Ganzalez A, Hartge P, Cerhan JR, et al. Body mass index and mortality among 1.46 million white adults. N Engl J Med. 2010; 363(23):2211–9.
  37. Blakely T, Foster R, Wilson N. Burden of Disease Epidemiology, Equity, and Cost-Effectiveness (BODE3) study protocol, version 2.1. Wellington, New Zealand: University of Otago; 2012. (Public health monograph series; vol. 30).
  38. Wilson R, Abbott JH. Age, period and cohort effects on body mass index in New Zealand, 1997–2038. Aust NZ J Public Health. 2018; 42:396–402.
  39. Abbott JH, Usiskin IM, Wilson R, et al. The quality-of-life burden of knee osteoarthritis in New Zealand adults: A model-based evaluation. PLOS ONE. 2017; 12(10):e0185676.
  40. GBD 2015 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990-2015: A systematic analysis for the Global Burde`n of Disease Study 2015. Lancet. 2016; 388:1545–602.
  41. Holt HL, Katz JN, Reichmann WM, et al. Forecasting the burden of advanced knee osteoarthritis over a 10-year period in a cohort of 60–64 year-old US adults. Osteoarthritis Cartilage. 2011; 19:44–50.
  42. Jordan JM, Helmick CG, Renner JB, et al. Prevalence of knee symptoms and radiographic and symptomatic knee osteoarthritis in African Americans and Caucasians: The Johnston County Osteoarthritis Project. J Rheumatol. 2007; 34:172–80.
  43. Cooper C, Snow S, McAlindon TE, et al. Risk factors for the incidence and progression of radiographic knee osteoarthritis. Arthritis Rheum. 2000; 43(5):995–1000.
  44. Brazier JE, Roberts J. The estimation of a preference-based measure of health from the SF-12. Med Care. 2004; 42(9):851–9.
  45. Pinto D, Robertson MC, Abbott JH, et al. Manual therapy, exercise therapy, or both, in addition to usual care, for osteoarthritis of the hip or knee. 2: Economic evaluation alongside a randomised controlled trial. Osteoarthritis Cartilage. 2013; 21:1504–13.
  46. Pharmaceutical Management Agency (PHARMAC). Cost Resource Manual, version 2.2. Wellington, New Zealand: PHARMAC; 2015.

For the PDF of this article,
contact nzmj@nzma.org.nz

View Article PDF

Osteoarthritis (OA) is a common and debilitating chronic disease and one of the leading causes of disability in New Zealand and worldwide.1–3 Knee OA is the most common form of OA, affecting as much as one-third of the population at some point in their lifetime. Globally, OA is responsible for more than 16 million disability-adjusted life years (DALYs) and is the 12th leading cause of worldwide disability.4 In New Zealand, osteoarthritis accounted for more than 15,000 DALYs in 2018.5 In addition to the health losses associated with OA, it also accounts for a substantial economic burden on the health system and the wider economy. In New Zealand, the total annual financial cost of arthritis (of all types) is estimated to be $4.2 billion, including health system costs of $990 million; productivity costs of $1.2 billion, through reduced employment, time off work and impaired performance at work; and other costs, such as non-health sector care and aids and modifications to support independent living for people with arthritis, of $2 billion.5

Total knee replacement (TKR) surgery is a common and successful operation to reduce pain and improve HRQoL in patients with advanced knee OA.6,7 Provision of TKR has been increasing across the world in recent decades.8–11 However, TKR is a costly procedure and high provision rates place strain on limited public healthcare resources. Furthermore, access to TKR, particularly in the public healthcare system, is limited by capacity constraints due to the availability of a suitably-trained workforce and surgical facilities.

The two most important risk factors for incidence and progression of knee OA are age and obesity. In New Zealand, as elsewhere, the combination of an ageing population and increasing rates of obesity is therefore expected to result in continuing increases in demand for OA healthcare.8,10,12–15 The population aged over 65 is expected grow by 40 percent over the next decade,16 while obesity rates are projected to increase to close to 50% of the adult population by the late 2030s.17 Understanding the implications of these population changes is critical for successful health system and workforce planning.

The aims of this study are (1) to estimate the projected healthcare costs of osteoarthritis care in New Zealand for the period 2013–2038, (2) to estimate the demand for TKR surgery in New Zealand over the same period, and (3) to assess the contribution of projected increases in population obesity to future healthcare expenditure and TKR demand.

Methods

We used the New Zealand Management of Osteoarthritis (NZ-MOA) model, a state-transition microsimulation model of the incidence, progression, health impact and healthcare costs of knee OA in New Zealand.18 The model generates a hypothetical cohort drawn from the 2013 New Zealand population distribution; runs each simulated individual through a sequence of annual health state transitions, defined by obesity, knee OA status (Kellgren-Lawrence grade), health-related quality of life (HRQoL) outcomes and treatment pathway; and computes the resulting per-capita healthcare costs, quality-adjusted life years (QALYs) and treatment utilisation. By running the same cohort through multiple scenarios, the incremental change in outcomes attributable to knee OA prevalence, change in risk factors or change in treatment patterns can be estimated.

We computed model outcomes under two scenarios: continuation of projected trends in population obesity levels17 and holding population obesity constant at 2013 levels. We generated simulation cohorts of 10,000 individuals for each age- (in five-year age groups), gender- and ethnicity-specific population subgroup. For each subgroup, we ran each scenario for a modelled 25-year time horizon (2013–2038) and recorded the age-, gender- and ethnicity-specific OA-related healthcare expenditure and TKR incidence in each year. Projections of healthcare costs and joint replacement provision at the population level were calculated by multiplying the age-, gender- and ethnicity-specific per-person projected outcomes from the simulation model by Statistics New Zealand’s published national population projections for each future period.19 This whole process was repeated 1,000 times, each with a new random draw from the distribution of obesity projections, resulting in a total simulated population of 760 million individuals to provide stable and precise estimates of the outcomes. We calculated the mean healthcare costs and joint replacement incidence in each year across the 1,000 cycles to provide the projected point estimates for each outcome, and the 2.5th and 97.5th percentiles to construct 95% uncertainty intervals for the contribution of increasing obesity levels (ie, undertaking probabilistic sensitivity analysis).

Baseline population model input parameters and sources have been described previously,18 and are summarised in Appendix A. Costs of treatment were based on provision of usual publicly and privately provided medical care as practiced in New Zealand,20 valued with New Zealand-specific reference prices21 (see Appendix Table A5). TKR incidence rates in the New Zealand population were obtained from the New Zealand Joint Registry,22 which covers almost all privately and publicly funded joint replacement surgeries in New Zealand, and used to calibrate the baseline modelled TKR provision rates. Projected increases in population obesity rates, stratified by age, gender and ethnicity, were obtained from a previously published age-period-cohort model of body mass index (BMI) in the New Zealand population.17

Cross-model validation was conducted by comparing our projected increases in the provision of TKR with those previously published in the Journal by Gary Hooper and colleagues (2014).13 Those estimates were based on different inclusion criteria than our model: all TKR were included for any primary diagnosis, whereas our model includes only one OA knee per person (ie, only the first primary TKR performed), and includes only TKR performed for a primary diagnosis of knee OA. The total number of TKR performed annually should therefore be lower in our model projections; to allow direct cross-model comparison, we normalised all estimates to an index value of 100 in 2011 (the base year in Hooper et al’s projections), to compare the total percentage increase in TKR provision between models. (The observed increase in TKR provision between 2011 and 2013 was used to determine the baseline (2013) index value for the NZ-MOA model, as the model was calibrated to observed 2013 data.)

Results

In 2013 (the base year for the NZ-MOA model), the total direct healthcare costs of knee OA were estimated to be NZ$199 million (1NZD≈0.82USD), and 5,070 patients had a first total joint replacement for knee OA. Women accounted for $110 million (55%) of the total costs and 2,460 TKR (48%); Māori for $18 million (9%) of the total costs and 490 TKR (10%).

The healthcare costs associated with OA treatment were projected to increase to $370 million per year in 2038 (at constant 2013 prices; Figure 1), an increase of 85%. Per-capita treatment costs (for the adult population aged 25 and over) were projected to increase from $69 to $90. TKR incidence was projected to increase to 9,040 per year in 2038 (Figure 2). The provision rate was projected to increase from 174 to 221 per 100,000 population per year.

Figure 1: Projected healthcare costs of knee osteoarthritis in New Zealand, 2013–2038.

c

Source: NZ-MOA simulation model. Lines show the projected increase in annual healthcare expenditure on OA treatment under the base case scenario (ongoing increase in population BMI based on current trends) and with no further increase in population BMI (ie, maintaining 2013 BMI distribution), relative to 2013 baseline level. ‘Attributable to BMI increase’ is the difference between the two lines, representing the incremental OA-related healthcare expenditure attributable to continued increases in population BMI relative to the 2013 level. The shaded uncertainty interval represents the 95% confidence interval for projected annual increase in population BMI. Right-hand axis shows total projected annual healthcare expenditure in each period for the base case and no BMI increase scenarios.
BMI indicates Body Mass Index; OA, Knee osteoarthritis.

Figure 2: Projected demand for first total knee replacement for knee osteoarthritis in New Zealand, 2013–2038.

c

Source: NZ-MOA simulation model. Lines show the projected increase in annual provision of TKR under the base case scenario (ongoing increase in population BMI based on current trends) and with no further increase in population BMI (ie, maintaining 2013 BMI distribution), relative to 2013 baseline level. ‘Attributable to BMI increase’ is the difference between the two lines, representing the incremental TKR provision attributable to continued increases in population BMI relative to the 2013 level. The shaded uncertainty interval represents the 95% confidence interval for projected annual increase in population BMI. Right-hand axis shows total projected annual provision in each period for the base case and no BMI increase scenarios.
BMI indicates Body Mass Index; TKR, Total knee replacement surgery.

In 2038, women were projected to account for $210 million (57%) of the total healthcare costs and 4,760 TKR (53%); Māori for $44 million (12%) of the total costs and 1,120 TKR (12%). Full results for projected costs and TKR incidence for each year 2013–2038, including ethnic and gender breakdowns, are reported in Tables B1 and B2 in the Appendix.

Further increases in population BMI, relative to the baseline distribution in 2013, accounted for 25% of the projected increase in per-capita OA-related healthcare costs and 47% of the increase in TKR provision rates between 2013 and 2038. By 2038, this accounts for an additional $22 million and 880 TKR per year. Over the 25-year time horizon, projected increases in BMI, relative to baseline levels, will result in an additional $231 million in OA healthcare costs and 9,880 TKR required.

The increase in TKR provision projected by our model, in the absence of further increases in population obesity, was very close to that previously reported, using a different modelling approach, by Hooper et al (2014) (Figure 3).13 Between 2011 and 2026, TKR provision was projected to increase by 48.8%, compared to 49.0% reported in the previously-published estimates. Allowing for projected increases in population obesity, this increased to 56.6% in our model, and 90.8% by 2038. The total numbers of projected TKR (without normalisation) from both models are shown in Figure B1 in the Appendix.

Figure 3: Observed and projected demand for total knee replacement in New Zealand, 1999–2038, cross-model comparison.

c

Source: NZ-MOA simulation model; NZJR 2016;22 Hooper et al 2014.13 Points show observed provision of TKR 1999–2013; dashed line the published projections from Hooper et al 2011–2026; and solid lines the new model projections 2013–2038, under the base case scenario (ongoing increase in population BMI based on current trends) and with no further increase in population BMI (ie, maintaining 2013 BMI distribution). The absolute numbers in each series differ due to different inclusion criteria: all values are normalised to 2011 = 100 for consistent comparisons.
BMI indicates Body Mass Index, in kg/m2; NZJR, New Zealand Joint Registry; TKR, Total knee replacement surgery.

Discussion

Treatment of knee OA cost the New Zealand healthcare system $200 million in 2013, an amount projected to increase to $370 million over 25 years. Demand for TKR is projected to increase by almost 80% over the same period, requiring an additional 4,000 surgeries to be performed annually by 2038. After adjusting for population growth, projected increases in obesity rates account for one-quarter of the increase in per-capita costs and almost half of the increase in TKR provision rates over 25 years.

These results highlight the need for effective—and cost-effective—treatment of OA throughout the disease course. Our projections were calculated under the assumption that patients’ access to TKR (and other treatments) remains at the same level as in 2013. Access to TKR is already rationed through the public health system, with many local health systems unable to offer surgery to all patients who would benefit;23,24 improving access to TKR would require further resourcing beyond that suggested by our projections. Effective, low-cost, early interventions, such as exercise therapy, can alleviate symptoms, improve patients’ quality of life and reduce the need for costly treatment, such as TKR, later in the disease course.25,26 Improving access to such treatments may help to mitigate the increasing burden on the health system associated with rising rates of OA.

There is limited existing research on the future healthcare costs of OA. Our projected increases in costs are very similar to those reported recently for Australia,27 which had a projected increase in direct OA healthcare costs of 2.2% per year between 2015 and 2030, based on the changing age and gender distribution of the population (ie, excluding the effect of changes in population obesity). For comparison, we projected the same 2.2% annual increase between 2013 and 2038 when holding obesity rates constant at the 2013 level. A study in Canada projected increases in costs of OA treatment of 4.7% per year between 2010 and 2031;28 these higher rates are at least in part explained by differing assumptions about the course of future healthcare prices.

There is more literature on projected provision of TKR, although previous studies have reported widely-varying results due to differences in context, modelling assumptions and statistical methodology. Our projections (2.3% per year, or 1.9% excluding the effects of increasing obesity, which have often not been captured in past studies) are at the low end of published projections for the US (between 1.9% and 8.5%14,29), similar to those for the UK (1.6% to 2.8%30) and the Netherlands (1.7% to 2.7%32), and higher than those for Sweden (0.9%8). In the local context, our results are consistent with those of Hooper et al,13 when excluding the effect of increasing obesity rates, providing external validation of our model, and demonstrate the additional burden being placed on the health system by the continuing obesity epidemic.

The study has limitations relating to model structure and availability of data. The estimated costs associated with knee OA are not stratified by disease severity (other than the cost of TKR for end-stage OA), due to a lack of available data on the relationship between disease progression and treatment costs. As obesity is associated with both incidence and progression of knee OA, this limitation may have resulted in our projections underestimating the effects of increasing obesity on healthcare costs of OA. We have also modelled direct healthcare costs only, excluding non-health costs such as time off work or reduced productivity, informal care outside the health system, or equipment and aids to assist with daily living. These other costs may be substantial in OA; for example, a recent report on the costs of (all types of) arthritis in New Zealand found that productivity losses associated with arthritis were 125% of direct healthcare costs, and other non-health financial costs a further 202% of healthcare costs.5 These suggest that the healthcare costs reported here may represent only one-quarter of the total societal cost of knee OA. Lastly, the NZ-MOA model captures the most severely OA-affected knee only for each individual; the projections of TKR provision therefore relate only to the first TKR surgery per patient. According to our data from the New Zealand Joint Registry, these account for approximately 75% of all TKR in New Zealand.

Strengths of the study include the use of a validated, state-of-the-art computer simulation model, populated with comprehensive and reliable national-level data on disease prevalence, risk factors and TKR provision.18 By modelling the underlying drivers of future TKR incidence—initial prevalence of OA across the population, and ongoing disease incidence, structural progression and symptom severity—this approach may provide more reliable projections than statistical models relying on linear or exponential extrapolation of observed trends in TKR incidence rates far beyond the period of observed data. The sensitivity of data-driven statistical projections of TKR incidence to modelling assumptions about the ongoing trend is apparent in the wide variability in previously published estimates. Our projected increases in TKR incidence were generally higher than previously published estimates assuming constant incidence rates (within demographic strata), but (often substantially) lower than those assuming the continuation of observed short-run increasing trends.

The consistent results found in our cross-model validation exercise demonstrate the validity of the model for predicting future outcomes based on current practice patterns and risk factor distributions and reliable estimates from Statistics New Zealand of projected demographic change. The impact of increasing population obesity, additional to these validated projections based on demographic change, was derived from published estimates, for the New Zealand population, of future trends in population BMI, and international evidence on the relationship between BMI and OA incidence and progression. By combining these reliable sources of data, using our well-validated computer simulation model, we have been able to provide well-grounded, coherent and reliable projections of the increasing healthcare burden of knee OA in the New Zealand population.

Conclusion

The healthcare burden of knee OA in New Zealand will continue to grow over the next 25 years due to population ageing and increasing rates of obesity. Without changes in the provision of effective and cost-effective care throughout the disease course, the annual direct healthcare costs of knee OA will increase by 85% to $370 million by 2038, and an additional 4,000 TKR surgeries per year will be required.

Appendix A

This appendix describes the sources and derivations of the NZ-MOA model input parameters used in this study. For further details refer to Wilson & Abbott (2018)33 and the NZ-MOA Technical Manual, version 1.4.0 (available on request from the authors).

Demographic characteristics

Baseline age, gender and ethnicity are drawn from the discrete joint probability distribution of the New Zealand population, obtained from the 2013 New Zealand Census.34 The model uses an annual cycle, so age is incremented by one year each period; gender and ethnicity are assumed to be fixed throughout the lifetime of each individual. Mortality rates, stratified by age, gender and ethnicity, were obtained from the New Zealand Period Life Tables 2012–14,35 and further adjusted for the relative risk of mortality associated with obesity using data from the US population.36 Mortality rates were assumed to decrease by 1.75% per year for non-Māori and 2.25% per year for Māori until 2026 and remain constant thereafter.37

Body mass index

Baseline BMI is drawn from an age-, gender- and ethnicity-specific log-normal distribution (Table A1). Future trajectories of BMI are assumed to follow the population-average age trajectory of BMI, plus an annual trend increase reflecting increases in population obesity.38 The ongoing trend increase was drawn, in each loop of the PSA process, from a normal distribution with gender- and ethnicity-specific mean and variance (Table A2).

Table A1: Mean of population baseline BMI distribution, by age, gender and ethnicity.

Table A2: Trend increase in population BMI distribution.

Structural (radiographic) knee osteoarthritis

Baseline radiographic knee OA status is drawn from the age-, gender-, ethnicity- and obesity-specific prevalence of self-reported doctor-diagnosed all-site OA, derived from the New Zealand Health Survey 2013/14 (Table A3), and further adjusted for the ratio of knee OA to all-site OA39 and the ratio of radiographically-defined knee OA to self-reported diagnosis.40 Incidence was derived to be consistent with baseline prevalence rates, assuming no case fatality or remission in radiographic knee OA.

Table A3: Baseline prevalence of self-reported doctor-diagnosed all-site OA.


Values are coefficients from a logistic regression model (ie, log(Odds(OA))). Further adjustments to estimate radiographically-defined knee OA prevalence are: ratio of knee to all-site OA (men = 0.713, women = 0.665), odds ratio of radiographically-defined knee OA to self-reported diagnosis (1.30).

Annual progression of radiographic knee OA (defined by K-L grade 2/3/4) was obtained from US- and UK-based prospective cohort studies,41–43 stratified by gender, obesity and prior K-L grade (Table A4).

Health-related quality of life impacts of radiographic knee osteoarthritis

The HRQoL impacts of knee OA are modelled on each of the six dimensions of the SF-6D, which are then valued using the SF-6D preference-based utility scores.44 The average health utility loss associated with knee OA is 0.02 at K-L grade 2, 0.05 at K-L grade 3, and 0.10 at grade 4.

OA healthcare costs

The direct healthcare costs of OA treatment were based on the assumed provision of usual medical care as practiced in New Zealand (consisting of GP consultations, analgesic medication, and referrals to physical therapy for some patients),45 valued with New Zealand-specific reference prices sourced from PHARMAC’s Cost Resource Manual for HTA in New Zealand (Table A5).46

Table A5: Annual costs of OA-related healthcare.

Appendix B

Table B1: Projected healthcare costs and TKR provision, by year.

All values refer to the New Zealand adult population aged 25–99.

Table B2: Projected healthcare costs and TKR provision, by year, sex and ethnicity.

All values refer to the New Zealand adult population aged 25–99.

Figure B1: Observed and projected demand for total knee replacement in New Zealand, 1999–2038, cross-model comparison.

c

Source: NZ-MOA simulation model; NZJR 2016;22 Hooper et al 2014.13
Points show observed provision of TKR 1999–2013; dashed line the published projections from Hooper et al 2001–2026; and solid lines the new model projections 2013–2038, under the base case scenario (ongoing increase in population BMI based on current trends) and with no further increase in population BMI (ie, maintaining 2013 BMI distribution).
Inclusion criteria differ between data sources, resulting in the discontinuity between observed and modelled TKR provision in 2013. See the main manuscript for further discussion of modelling assumptions and inclusion criteria.
BMI indicates Body Mass Index, in kg/m2; NZJR, New Zealand Joint Registry; TKR, Total knee replacement surgery.  

Summary

Abstract

Aim

To estimate the healthcare costs and demand for total knee replacement (TKR) associated with knee osteoarthritis in New Zealand over the period 2013-2038 and the contribution of increasing obesity rates to these costs.

Method

We used the NZ-MOA computer simulation model of knee osteoarthritis in the New Zealand population. Osteoarthritis-related healthcare costs and TKR incidence were modelled for a 25-year horizon, for a starting cohort drawn from the 2013 New Zealand population. Population obesity projections were used to estimate the life-course of cohort obesity. Per-person projected outcomes were multiplied by national demographic population projections to obtain total population projections.

Results

Healthcare costs of knee osteoarthritis were projected to increase from NZ$199 million in 2013 to $370 million in 2038. Annual TKR incidence was projected to increase from 5,070 to 9,040 over the same period. Projected increases in population obesity rates (above the obesity prevalence seen in 2013) accounted for 25% and 47% of the projected increase in per-capita healthcare costs and TKR provision rates, respectively.

Conclusion

The healthcare burden of knee OA will continue to increase for the foreseeable future. Public health measures to reduce further increases in population obesity rates would contribute to slowing this rising burden.

Author Information

Ross Wilson, Research Fellow, Centre for Musculoskeletal Outcomes Research, University of Otago, Dunedin; J Haxby Abbott, Research Professor, Centre for Musculoskeletal Outcomes Research, University of Otago, Dunedin.

Acknowledgements

Correspondence

Dr Ross Wilson, Centre for Musculoskeletal Outcomes Research, Department of Surgical Sciences, University of Otago, PO Box 56, Dunedin 9054.

Correspondence Email

ross.wilson@otago.ac.nz

Competing Interests

Dr Wilson reports grants from Health Research Council of New Zealand during the conduct of the study.

  1. Cross M, Smith E, Hoy D, et al. The global burden of hip and knee osteoarthritis: Estimates from the Global Burden of Disease 2010 study. Ann Rheum Dis. 2014; 73:1323–30.
  2. Ministry of Health, Health loss in New Zealand 1990-2013: A report from the New Zealand Burden of Diseases, Injuries and Risk Factor Study. Wellington, New Zealand: Ministry of Health; 2016.
  3. GBD 2015 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990-2015: A systematic analysis for the Global Burden of Disease Study 2015. Lancet. 2016; 388:1545–602.
  4. Institute for Health Metrics and Evaluation. GBD Compare. Available at: http://gbd2016.healthdata.org/gbd-compare/ (accessed 2019/04/17); 2019.
  5. Deloitte Access Economics. The economic cost of arthritis in New Zealand in 2018. Arthritis New Zealand; 2018 Aug.
  6. Bruyère O, Ethgen O, Neuprez A, et al. Health-related quality of life after total knee or hip replacement for osteoarthritis: A 7-year prospective study. Arch Orthop Trauma Surg. 2012; 132:1583–7.
  7. Ethgen O, Bruyère O, Richy F, et al. Health-related quality of life in total hip and total knee arthroplasty: A qualitative and systematic review of the literature. J Bone Joint Surg. 2004; 86-A:963–74.
  8. Robertsson O, Dunbar MJ, Knutson K, Lidgren L. Past incidence and future demand for knee arthroplasty in Sweden: A report from the Swedish Knee Arthroplasty Register regarding the effect of past and future population changes on the number of arthroplasties performed. Acta Orthop Scand. 2000; 71:376–80.
  9. Kim H-A, Kim S, Seo YI, et al. The epidemiology of total knee replacement in South Korea: National registry data. Rheumatology. 2008; 47:88–91.
  10. Pedersen AB, Johnsen SP, Overgaard S, et al. Total hip arthroplasty in Denmark: Incidence of primary operations and revisions during 1996–2002 and estimated future demands. Acta Orthop. 2005; 76:182–9.
  11. Ravi B, Croxford R, Reichmann WM, et al. The changing demographics of total joint arthroplasty recipients in the United States and Ontario from 2001 to 2007. Best Pract Res Clin Rheumatol. 2012; 26:637–47.
  12. Kurtz SM, Lau E, Ong K, et al. Future young patient demand for primary and revision joint replacement: National projections from 2010 to 2030. Clin Orthop Relat Res. 2009; 467:2606–12.
  13. Hooper G, Lee AJ-J, Rothwell A, Frampton C. Current trends and projections in the utilisation rates of hip and knee replacement in New Zealand from 2001 to 2026. NZ Med J. 2014; 127:82–93.
  14. Kurtz S, Ong K, Lau E, et al. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am. 2007; 89-A:780–5.
  15. Nemes S, Gordon M, Rogmark C, Rolfson O. Projections of total hip replacement in Sweden from 2013 to 2030. Acta Orthop. 2014; 85:238–43.
  16. Statistics New Zealand. National population projections: 2016(Base)–2068. Available at: http://www.stats.govt.nz/information-releases/national-population-projections-2016base2068 (accessed 2019/04/04); 2016.
  17. Wilson R, Abbott JH. Age, period and cohort effects on body mass index in New Zealand, 1997–2038. Aust NZ J Public Health. 2018; 42:396–402.
  18. Wilson R, Abbott JH. Development and validation of a new population-based simulation model of osteoarthritis in New Zealand. Osteoarthritis Cartilage. 2018; 26:531–9.
  19. Statistics New Zealand. National ethnic population projections: 2013(Base)–2038 (update). Available at: http://www.stats.govt.nz/information-releases/national-ethnic-population-projections-2013base2038-update (accessed 2018/11/22); 2017.
  20. Pinto D, Robertson MC, Abbott JH, et al. Manual therapy, exercise therapy, or both, in addition to usual care, for osteoarthritis of the hip or knee. 2: Economic evaluation alongside a randomized controlled trial. Osteoarthritis Cartilage. 2013; 21:1504–13.
  21. PHARMAC, Cost Resource Manual version 2.2. PHARMAC; 2015.
  22. New Zealand Joint Registry, Sixteen year report: January 1999 to December 2015. New Zealand Joint Registry; 2016 Oct.
  23. Gwynne-Jones D. Quantifying the demand for hip and knee replacement in Otago, New Zealand. NZ Med J. 2013; 126:7–17.
  24. Harcombe H, Davie G, Derrett S, et al. Equity of publicly-funded hip and knee joint replacement surgery in New Zealand: Results from a national observational study. NZ Med J. 2016; 129:8–18.
  25. Abbott JH, Wilson R, Pinto D, et al. Incremental clinical effectiveness and cost effectiveness of providing supervised physiotherapy in addition to usual medical care in patients with osteoarthritis of the hip or knee: 2-year results of the MOA randomised controlled trial. Osteoarthritis Cartilage. 2019; 27:424–34.
  26. Teoh LSG, Eyles JP, Makovey J, et al. Observational study of the impact of an individualized multidisciplinary chronic care program for hip and knee osteoarthritis treatment on willingness for surgery. Int J Rheum Dis. 2017; 20:1383–92.
  27. Ackerman IN, Pratt C, Gorelik A, Liew D. Projected burden of osteoarthritis and rheumatoid arthritis in Australia: A population-level analysis. Arthritis Care Res. 2018; 70:877–83.
  28. Sharif B, Kopec J, Bansback N, et al. Projecting the direct cost burden of osteoarthritis in Canada using a microsimulation model. Osteoarthritis Cartilage. 2015; 23:1654–63.
  29. Inacio MCS, Paxton EW, Graves SE, et al. Projected increase in total knee arthroplasty in the United States – an alternative projection model. Osteoarthritis Cartilage. 2017; 25:1797–803.
  30. Culliford D, Maskell J, Judge A, et al. Future projections of total hip and knee arthroplasty in the UK: Results from the UK Clinical Practice Research Datalink. Osteoarthritis Cartilage. 2015; 23:594–600.
  31. Patel A, Pavlou G, Mújica-Mota RE, Toms AD. The epidemiology of revision total knee and hip arthroplasty in England and Wales: A comparative analysis with projections for the United States. A study using the National Joint Registry dataset. Bone Joint J. 2015; 97-B:1076–81.
  32. Otten R, Roermund PM van, Picavet HS. Trends in the number of knee and hip arthoplasties: Considerably more knee and hip prostheses due to osteoarthritis in 2030. Ned Tijdschr Geneeskd. 2010; 154:A1534.
  33. Wilson R, Abbott JH. Development and validation of a new population-based simulation model of osteoarthritis in New Zealand. Osteoarthritis Cartilage. 2018; 26:531–9.
  34. Statistics New Zealand. 2013 Census. Wellington, New Zealand: Statistics New Zealand; 2014.
  35. Statistics New Zealand. New Zealand period life tables, 2012–2014. Wellington, New Zealand: Statistics New Zealand; 2015.
  36. Berrington de Ganzalez A, Hartge P, Cerhan JR, et al. Body mass index and mortality among 1.46 million white adults. N Engl J Med. 2010; 363(23):2211–9.
  37. Blakely T, Foster R, Wilson N. Burden of Disease Epidemiology, Equity, and Cost-Effectiveness (BODE3) study protocol, version 2.1. Wellington, New Zealand: University of Otago; 2012. (Public health monograph series; vol. 30).
  38. Wilson R, Abbott JH. Age, period and cohort effects on body mass index in New Zealand, 1997–2038. Aust NZ J Public Health. 2018; 42:396–402.
  39. Abbott JH, Usiskin IM, Wilson R, et al. The quality-of-life burden of knee osteoarthritis in New Zealand adults: A model-based evaluation. PLOS ONE. 2017; 12(10):e0185676.
  40. GBD 2015 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990-2015: A systematic analysis for the Global Burde`n of Disease Study 2015. Lancet. 2016; 388:1545–602.
  41. Holt HL, Katz JN, Reichmann WM, et al. Forecasting the burden of advanced knee osteoarthritis over a 10-year period in a cohort of 60–64 year-old US adults. Osteoarthritis Cartilage. 2011; 19:44–50.
  42. Jordan JM, Helmick CG, Renner JB, et al. Prevalence of knee symptoms and radiographic and symptomatic knee osteoarthritis in African Americans and Caucasians: The Johnston County Osteoarthritis Project. J Rheumatol. 2007; 34:172–80.
  43. Cooper C, Snow S, McAlindon TE, et al. Risk factors for the incidence and progression of radiographic knee osteoarthritis. Arthritis Rheum. 2000; 43(5):995–1000.
  44. Brazier JE, Roberts J. The estimation of a preference-based measure of health from the SF-12. Med Care. 2004; 42(9):851–9.
  45. Pinto D, Robertson MC, Abbott JH, et al. Manual therapy, exercise therapy, or both, in addition to usual care, for osteoarthritis of the hip or knee. 2: Economic evaluation alongside a randomised controlled trial. Osteoarthritis Cartilage. 2013; 21:1504–13.
  46. Pharmaceutical Management Agency (PHARMAC). Cost Resource Manual, version 2.2. Wellington, New Zealand: PHARMAC; 2015.

Contact diana@nzma.org.nz
for the PDF of this article

Subscriber Content

The full contents of this pages only available to subscribers.

LOGINSUBSCRIBE