ABSTRACT
Background This study investigated the cost-effectiveness of vitamin D3 supplementation in older adults in Ireland, with year-round vitamin D deficiency (serum 25-hydroxyvitamin D concentration <30 nmol/L) (13% of Irish adults), from the perspective of the Health Service Executive (HSE).
Methods Three age groups were investigated: (1) ≥50 years, (2) ≥60 years, (3) ≥70 years. Based on the clinical literature, vitamin D3 supplementation may (1) decrease all-cause mortality by 7%, and (2) reduce hip fractures by 16%, and non-hip fractures by 20%. A discount rate of 4% was applied to life years and QALYs gained. The annual healthcare costs per patient used in the model are based on the average annual health resource use over the 5-year time horizon of the model.
Results The cost/QALY estimates in all three age groups are below the usually acceptable cost-effectiveness threshold of €20,000/QALY. The most cost-effective and least costly intervention was in adults ≥70 years. For this age group, the average annual costs and outcomes would be approx. €6.2 million, 1,043 QALYs gained, with a cost/QALY of approx. €6,000. The results are most sensitive to the mortality risk reduction following vitamin D3 supplementation.
Conclusion The cost-effectiveness of vitamin D3 supplementation is most robust in adults ≥70 years. Clinical uncertainty in the magnitude of the benefits of vitamin D3 supplementation could be further addressed by means of (1) performing a clinical research study or (2) conducting a pilot/regional study, prior to reaching a decision to invest in a full nationwide programme.
INTRODUCTION
Clinical vitamin D deficiency (serum 25-hydroxyvitamin D (25(OH)D) concentration below <30 nmol/L) increases the risk of excess mortality and disease [1]. Vitamin D3 supplementation is likely to be clinically most beneficial in deficiency [1]. Vitamin D deficiency, as measured by serum 25(OH)D, is particularly high amongst older Irish adults [2].
There is a growing interest in the cost-effectiveness of vitamin D to prevent disease. While several previous studies were mainly conducted in the elderly for fall and/or fractures (e.g., [3,4]), more recently, a study estimated the costs and savings for preventing cancer deaths by vitamin D supplementation of the population aged ≥50 years in Germany [5]. The results of this study supported the use of vitamin D3 supplementation among older adults as a potential cost saving approach to substantially reduce cancer mortality [5].
The objective of this study is to investigate the cost-effectiveness of vitamin D3 supplementation in older adults in Ireland, with year-round vitamin D deficiency (25(OH)D concentration <30 nmol/L) (13% of adults [2]), from the perspective of the Health Service Executive (HSE).
METHODS
Deaths and fractures avoided
The methodology used builds upon that used by Niedermaier, et al. [5]. The methodology applied to Ireland uses Central Statistics Office (CSO) Irish Life Table No. 17, for males and females, for the period, 2015-2017 [6]. The data for male and females from the Life Table were combined using Irish population statistics for 2016 [7] and applying the method of weighted averaging. The annual mortality probability and the expected life expectancy were determined for all ages. A 4% annual discount rate [8] was applied in order to calculate the discounted life expectancy. The “at risk” population of interest (i.e., those with year-round vitamin D deficiency) is estimated to be 13% of the Irish population [2]. This population cohort was estimated by applying 13% to the Irish population statistics for 2016 for all ages. The population data were summarised into 5-year age groups, beginning at age 50 years. The final age group was for those ≥85 years (with a mean age assumed to be 90 years, when required). For any given 5-year age group, the expected annual number of deaths is equal to the number of people in the age group multiplied by the annual mortality probability for the mid-point age of the group.
Based on the clinical literature, vitamin D3 supplementation may (1) decrease all-cause mortality by 7% [9], of which 4.2% (of the 7%) is a reduction in cancer mortality, and (2) reduce hip fractures by 16% (and related excess mortality) [10], and non-hip fractures by approx. 20% [11,12]. In the base-case analysis, the reduction in all-cause mortality was implemented in the model by applying a 7% reduction to the expected number of deaths occurring in each 5-year age group. This point-estimate reduction in all-cause mortality was varied over the range 1% to 10% in the univariate sensitivity analysis.
The number of hip fractures in 2019 and their distribution by age were obtained from the Hip Fracture database [13]. The number of non-hip fractures in 2014 was estimated from the total number of fractures and the number of hip fractures reported in the literature [14]. The number of non-hip fractures was assumed to have the same relative age distribution, as was found for hip fractures. It was conservatively assumed that only 13% of people who have a fracture would have year-round vitamin D deficiency and be included in the “at risk” population of interest [2]. The reductions in hip fractures (16%, and the associated estimated 1-year mortality of 22% following a hip fracture [15]) and non-hip fractures (approx. 20%) were included in the base-case analysis and varied in the univariate sensitivity analysis.
The results obtained from the model were investigated for three age cohorts: (1) ≥50 years, (2) ≥60 years, (3) ≥70 years.
Discounted life years and discounted QALYs gained
The number of discounted life years gained for any given age group is equal to the number of deaths avoided multiplied by the associated discounted life expectancy. The number of discounted QALYs gained for any given age group is equal to number of discounted life years gained multiplied by the age-related utility value for the mid-point of the age group.
EQ-5D index population norms for the UK [16] were used for age-related utility values. The following age-related utility equation was used, for ages ≥50 years (which had a coefficient of determination (R2) of approx. 99%).
The disutility estimate used for hip-fracture was 0.20 and for non-hip-fracture was 0.09 [4]. These disutility estimates were applied for one year to estimate fracture related QALYs gained, i.e., number of fractures avoided per annum multiplied by the relevant fracture disutility.
Comparator and time horizon
The comparator is the current standard of care, which is being compared against a systematic public health programme (1) to identify year-round vitamin D deficient adults (≥50 years), by means of serum 25(OH)D measurement, and (2) to treat such adults with vitamin D3 supplementation, under GP supervision, to render the patients vitamin D replete and to maintain their vitamin D adequacy over time.
A 5-year time horizon was employed in the sense that it was assumed that it would take 5-years for the public health programme to have identified the eligible vitamin D deficient patients and for them to have been treated for their vitamin D deficiency, and to be maintained in a vitamin D replete state.
Healthcare resource use and costs
For each age-group, it was assumed that the average annual healthcare cost would be the cost of treating all the eligible patients plus the cost of patient identification (i.e., a blood 25(OH) vitamin D measurement in order to identify the “at risk” population), with the latter “set-up” cost averaged over 5-years. Thus, only those patients who are GP-assessed to be year-round vitamin D deficient are subsequently treated with vitamin D3 supplementation and monitored. Vitamin D deficiency was assumed to be treated with vitamin D3 4,000 IU daily for up to 10 weeks, followed by 800 IU daily thereafter. Healthcare costs included in the analysis are: serum 25(OH)D measurement, the average cost of a GP visit, and drug acquisition costs of vitamin D3 supplements.
The cost off-sets included in the analysis were end-of-life invasive cancer costs avoided and the costs of fractures (hip and non-hip) avoided. The reduction in all-cause mortality was implemented in the model by applying a 7% reduction to the expected number of deaths occurring in each 5-year age group. This 7% reduction in all-cause mortality could be “partitioned” into a 4.2% reduction in cancer mortality and a 2.8% reduction in non-cancer mortality [9]. Thus, approx. 61% of the reduction in mortality was due to the reduction in cancer mortality. Invasive cancers constitute approx. 55% of all cancers [17]. Thus, approx. 33% of the reduction in mortality was due to the reduction in invasive cancer mortality. Conservatively, for deaths avoided, end-of-life costs of invasive cancers avoided were the only cost off-sets included in the analysis.
Average UK end-of-life cancer costs [18] were applied to Irish invasive cancer deaths, using the method of purchasing power parity (PPP), with inflation of healthcare costs to €(2020) values. Additional healthcare cost off-sets included in the analysis were due to the healthcare costs of hip and non-hip fractures avoided. The unit costs of all healthcare resource use items included in the model are summarised in Table 1.
The cost-effectiveness analysis did not include any discounting of healthcare costs.
Univariate and multi-variate sensitivity analyses
The univariate sensitivity analyses and multi-variate sensitivity analysis performed are summarised in Table 2.
All results of the data analysis given below were obtained using Microsoft Excel 2019, 32-bit version.
RESULTS
Basecase results
The basecase cost-effectiveness results are summarised in Table 3. The cost/QALY estimates in all three age groups are below the usually acceptable cost-effectiveness threshold of €20,000/QALY in Ireland. The most cost-effective and least costly intervention was in adults ≥70 years of age. While a public health primary prevention programme in adults ≥70 years of age would produce fewer QALYs, it would be sufficiently less costly to result in the lowest (best) cost/QALY.
All additional basecase results will be focused on those obtained in adults ≥70 years of age.
The annual healthcare cost offsets reduce the average annual costs from a total of approximately €9.0 million to approximately €6.2 million per annum for treating older adults ≥ 70 years, who are year-round vitamin D deficient, with vitamin D3 supplementation (Figure 1).
Abbreviations: vit D, vitamin D
Vitamin D3 supplementation drug acquisition costs are the biggest cost component, being approximately 41% of the total €9.0 million annual healthcare costs (excluding cost offsets) (Figure 2). Approximately 84% of the annual healthcare cost off-sets are the healthcare costs of fractures avoided, with the remaining approximate 16% of the annual healthcare cost off-sets due to end-of-life care costs avoided from a reduction in invasive cancer deaths.
Abbreviations: vit D, vitamin D
Results of the sensitivity analyses
The impact on the cost/QALY of varying the % reduction in all-cause mortality from treating older adults, who are year-round vitamin D deficient, with vitamin D3 supplementation is given in Figure 3. As can be seen, for elderly adults ≥70 years of age, even if the reduction in all-cause mortality due to Vitamin D3 supplementation was reduced to 2% (7% in the basecase), the cost/QALY in this age group would still be less than €20,000 per QALY gained. A similar threshold analysis can be performed for the other age-groups by visible inspection of Figure 3.
Abbreviations: wrt, with respect to; vit D, vitamin D; QALY, quality adjusted life year
The impact on the cost/QALY of varying the risk reduction of fractures (hip and non-hip) from treating older adults, who are year-round vitamin D deficient, with vitamin D3 supplementation was investigated; in all cases explored, the cost/QALY remain less than €20,000 per QALY gained, for each of the three age-groups (Figure 4).
Abbreviations: vit D, vitamin D; QALY, quality adjusted life year
The impact on the cost/QALY of varying the discount rate per annum (4% in the basecase) was investigated; in all cases explored, the cost/QALY remain less than €20,000 per QALY gained, for all age-groups (Figure 5).
Abbreviations: vit D, vitamin D; QALY, quality adjusted life year
In terms of healthcare costs, the most uncertain are in relation to (1) serum 25(OH)D measurement, and (2) average end-of-life invasive cancer cost. In a univariate sensitivity analysis:
The serum 25(OH)D measurement cost was increased by 50% (compared to the basecase, from €21.50 to €32.25), which will tend to increase the cost per QALY results (Table 3). While the cost per QALY results increased modestly for each age cohort compared to the basecase results, but remain below the €20,000/QALY threshold (Table 3).
The average end-of-life invasive cancer cost was increased by 50% (compared to the basecase, from €6,487 to €9,731), as it was based on a 2005 reference [17]. This cost off-set will tend to decrease the cost per QALY results (Table 3).
The cost per QALY results were not very sensitive to large changes (50% increase) in either of these uncertain unit costs.
Potentially, some vitamin D deficient elderly adults may require additional serum 25(OH)D measurement, GP visit(s) per annum, and/or additional course(s) of vitamin D3 supplementation to correct their vitamin D deficiency. To investigate this, a sensitivity analysis was conducted in which the average annual cost of serum 25(OH)D measurement, GP visit, plus vitamin D3 supplementation acquisition cost (€129.44 in the basecase) was increased by 50% (to €194.17) (Table 3). In this case, the cost per QALY results increased for each age cohort compared to the basecase results; however, for both the age groups: ≥60 years and ≥70 years, the cost per QALY were well below €20,000/QALY (Table 3).
Results of a “pessimistic” scenario
The parameterisation for a multi-variate sensitivity analysis performed is summarised in Table 2. In this case, the potential clinical benefits of vitamin D3 supplementation to vitamin D deficient elderly adults are decreased by 30% compared to the basecase, whereas, the relevant healthcare costs are increased by 30%. This is regarded as a “pessimistic” scenario.
In this “pessimistic” scenario, the cost per QALY for the age group: ≥70 years remained well below €20,000/QALY, whereas, this was no longer the case for the age group: ≥60 years, with the cost per QALY considerably above €20,000/QALY for the age group: ≥50 years (Table 3).
DISCUSSION
Vitamin D3 supplementation is likely to be clinically most beneficial in deficiency [1]. Furthermore, those most likely to gain maximum benefit from vitamin D3 supplementation are those, with year-round vitamin D deficiency (25(OH)D concentration <30 nmol/L), which is estimated to be 13% of Irish adults [2]. What has been evaluated in this paper is the healthcare costs and benefits of (1) identifying year-round vitamin deficient adults (≥50 years), by means of serum 25(OH)D measurement, and (2) treating such adults with vitamin D3 supplementation, under GP supervision, to render the patients vitamin D replete and to maintain their vitamin D adequacy over time. It is envisaged that GP supervision could be performed through routine GP / patient interaction. However, it was also assumed that those patients identified as having year-round vitamin D deficiency would require, on average, one additional GP visit per annum, with a further additional GP visit in the year that the qualifying patient would enter the potential public health programme. It is further assumed that serum 25(OH)D measurement during the summer the summer months is sufficient to identify those older adults with year-round vitamin D deficiency. No calcium supplements were included in the cost-effectiveness analysis, as it was assumed that adequate calcium intake by patients could be achieved through dietary advice provided as part of the GP visit.
The cost-effectiveness modelling assumed that patient identification and treatment with vitamin D3 supplementation would take some years to be achieved. Also, the benefits of treating vitamin D deficiency would take time to be manifested in terms of clinical benefits. Therefore, it was assumed that a “steady-state” would be achieved by the end of 5-years from initiation of the public health programme. For each age-group, it was assumed that the average annual healthcare cost would be the cost of treating all the eligible patients plus the cost of patient identification, with the latter “set-up” costs averaged over 5-years. It is unknown, however, whether this time frame would be sufficient to capture the potential benefits of vitamin D3 supplementation on reducing all-cause mortality or reducing the risk of fractures.
The cost/QALY estimates in all three age groups are below the usually acceptable cost-effectiveness threshold of €20,000/QALY in Ireland. The most cost-effective and least costly intervention was in adults ≥70 years of age. While a public health primary prevention programme in adults ≥70 years of age would produce fewer QALYs, it would be sufficiently less costly to result in the lowest (best) cost/QALY. Therefore, it is proposed that a GP-monitored, vitamin D3 supplementation public health programme be considered in adults ≥70 years of age, in the first instance.
The cost-effectiveness results could potentially be improved if additional clinical benefits of vitamin D3 supplementation had been included in the model, e.g., decreasing the incidence of respiratory tract infections and the adverse clinical consequences than can arise from such infections [24]. In addition, a conservative approach was taken where applicable. For example, the cost off-sets due to cancer end-of-life care were limited to invasive cancer care costs only, with invasive cancers constituting approx. 55% of all cancers [17]. In addition, discounting was applied to healthcare benefits (life years / QALYs gained from deaths avoided), but not to healthcare costs, such as the “set-up” costs, involving serum 25(OH)D measurement.
The results of the cost effectiveness analysis are most sensitive to the mortality risk reduction following vitamin D3 supplementation. For elderly adults ≥70 years of age, it was found that even if the reduction in all-cause mortality due to vitamin D3 supplementation was reduced to 2% (7% in the basecase), the cost/QALY in this age group would still be less than €20,000 per QALY gained. However, potentially the reduction in all-cause mortality due to vitamin D3 supplementation could be less than this.
Typically, the uncertainty in cost-effectiveness analysis can be assessed through probabilistic sensitivity analysis (PSA). A PSA was is not performed in this study. However, the nature of the clinical evidence base for vitamin D3 supplementation may not be very suitable for clinical uncertainty to be assessed in such a manner. The clinical uncertainty might be better addressed by means of (1) performing a clinical research study prior to reaching a decision to invest in a GP-monitored, vitamin D3 supplementation public health programme or (2) conducting a pilot/regional study prior to reaching a decision to invest in a full nationwide programme.
CONCLUSION
The cost/QALY estimates of a GP-monitored, vitamin D3 supplementation public health programme in the three age groups: (1) ≥50 years, (2) ≥60 years, and (3) ≥70 years, are below the usually acceptable cost-effectiveness threshold of €20,000/QALY in Ireland. The cost-effectiveness of vitamin D3 supplementation is most robust in adults ≥70 years. The results of the cost effectiveness analysis are most sensitive to the mortality risk reduction following vitamin D3 supplementation. It is proposed that a GP-monitored, vitamin D3 supplementation public health programme be considered in adults ≥70 years of age, in the first instance.
Data Availability
The methodology used is described in detail in the manuscript. The health economic analysis is available as a MS Excel spreadsheet
Contributors
All authors reviewed and verified the final manuscript for submission. All authors were involved in the design and interpretation of the economic evaluation, although the analysis was carried out by LFL.
Funding
No financial support was received for any aspect of this research.
Competing interests
None declared.
Patient consent for publication
Not applicable.
Ethics approval
Not applicable.
