Economic analysis of Heart and Stroke Foundation of Ontario’s Hypertension Management Initiative

Abstract

Objectives

Hypertension is suboptimally treated in primary care settings. We evaluated the cost-effectiveness of the Heart and Stroke Foundation of Ontario’s Hypertension Management Initiative (HMI), an interdisciplinary, evidence-informed chronic disease management model for primary care that focuses on improving blood pressure management and control by primary care providers and patients according to clinical best practice guidelines.

Methods

The perspective of our analysis was that of the Ontario Ministry of Health and Long-Term Care with a lifetime horizon and 5% annual discount rate. Using data from a prospective cohort study from the HMI, we created two matched groups: pre-HMI (standard care), and post-HMI (n = 1720). For each patient, we estimated the 10-year risk of cardiovascular disease (CVD) using the Framingham risk equation and life expectancy from life tables. Long-term health care costs incurred with physician visits, acute and chronic care hospitalizations, emergency department visits, same-day surgeries, and medication use were determined through linkage to administrative databases, using a bottom-up approach.

Results

The HMI intervention was associated with significant reductions in systolic blood pressure (126 mmHg vs 134 mmHg with standard care; P-value < 0.001). These improvements were associated with a reduction in the 10-year risk of CVD (9.5% risk vs 10.7% in standard care; P-value < 0.001) and a statistically significant improvement in discounted life expectancy (9.536 years vs 9.516 in standard care; P-value < 0.001). The HMI cohort had a discounted mean lifetime cost of $22,884 CAD vs $22,786 CAD for standard care, with an incremental cost-effectiveness ratio of $4939 CAD per life-year gained.

Conclusion

We found that the HMI is a cost-effective means of providing evidence-informed, chronic disease management in primary care to patients with hypertension.

Keywords:

• hypertension
• economic evaluation
• cardiovascular disease

Supplementary materials

1) 10-year cardiovascular disease (CVD) and component risks

For each patient, we determined the 10-year risk of CVD using the Framingham risk equation:

$\widehat{p}=1-S_0{(t)}^{\text{exp}\left({\sum }_{i=1}^p{\beta }_i{X}_i-{\sum }_{i=1}^p{\beta }_i{\overline{X}}_i\right)},$(S1)

where S₀(t) is the baseline survival at follow-up time t, β_i is the estimated regression coefficient (log hazard ratio; see Table 2 from D’Agostinho et al¹), X_i is the log-transformed value of the i^th risk factor (if continuous), X̄_i is the corresponding mean, and p denotes the number of risk factors.

Risk estimation from Cox model

Example

Women (baseline 10-year survival 0.95012): For a 61-year-old woman who has not been treated for high blood pressure, has a total cholesterol of 180 mg/dL, high-density lipoprotein of 47 mg/dL, systolic blood pressure of 124 mmHg, and is a current smoker but is not diabetic (see Table 11 from D’Agostinho et al¹), the risk estimate based on the Cox model is computed as follows:

$\begin{array}{l}\mathrm{\Sigma }{\beta }_{\text{i}}{\text{X}}_{\text{i}}=2.32888\times \text{log}(61)+1.20904\times \text{log}(180)-0.70833\\ \times \text{log}(47)+2.76157\times \text{log}(124)+2.82263\\ \times 0+0.52873\times 1+0.69154\times 0=26.9653\end{array}$(S2) $\begin{array}{l}\mathrm{\Sigma }{\beta }_{\text{i}}=2.32888\times 3.8686+1.20904\times 5.3504-0.70833\\ \times 4.0176+2.76157\times 4.2400+2.82263\times 0.5826\\ +0.52873\times 0.3423+0.69154\times 0.0376=26.1931\end{array}$(S3)

10-year risk of

$\begin{array}{l}\text{CVD}={\text{R}}_{\text{cvd}}=\widehat{p}=1-S_0{(t)}^{\text{exp}\left({\sum }_{i=1}^p{\beta }_i{X}_i-{\sum }_{i=1}^p{\beta }_i{\overline{X}}_i\right)},\\ =1-{0.95012}^{\text{exp}(26.9653-26.1931)}=0.1048\end{array}$(S4)

Therefore, using calibration factors (see Tables 3 and 4 from D’Agostinho et al¹), the 10-year risk of:

1. CHD = 0.1048 × 0.6086 = 0.06378,

2. stroke = 0.1048 × 0.2385 = 0.02499, and so forth.

Life expectancy

Assumption

On average, the CVD event will occur at the mid-point of the 10-year projection. Therefore, we will include 5 years on life-expectancy for CVD.

For each patient, the estimated life expectancy will be:

$({\text{R}}_{\text{cvd}}\times [\text{CVD}-\text{life}-\text{expectancy}+5(1-(\text{annual}-\text{non\hspace{0.17em}CVD\hspace{0.17em}mortality\hspace{0.17em}rate})])+([1-\text{Rcv}]\times \text{healthy}-\text{life}-\text{expectancy})$(S5)

Example

A 50-year-old-man has a R_cvd of 0.13. Therefore, using Table 4 from Peeters et al,

$\begin{array}{l}\text{Life\hspace{0.17em}expectancy}=({\text{R}}_{\text{cvd}}\times \text{CVD}-\text{life}-\text{expectancy})+([1-{\text{R}}_{\text{cvd}}]\times \text{healthy}-\text{life}-\text{expectancy})\\ =(0.13\times [15.9+5])+([1-0.13]\times 26.7)\\ =26.0\hspace{0.17em}\text{years}\end{array}$

2) Discounted life expectancy

To calculate the discounted life expectancy (DLE), the life expectancy of each participant from above will be assumed to be the area under the curve of an exponential survival curve (S(t) = e^−kt) for that particular patient.

Using the declining exponential approximation of life expectancy (DEALE) method, k can be calculated for each patient. Applying a continuous discounting rate, r, of 5%, the discounted LE is equal to the following:

$\begin{array}{l}\text{DLE}={\int }_0^{\infty }S(t)e^{-rt}dt\\ \text{DLE}={\int }_0^{\infty }{e}^{-kt}{e}^{-rt}dt={\int }_0^{\infty }{e}^{-kt-rt}dt\\ ={\int }_0^{\infty }{e}^{-(k+r)t}dt\frac{-1}{(k+r)}{\int }_0^{\infty }{e}^{-(k+r)t}dt\\ =\frac{-1}{(k+r)}(e^{-(k+r)\infty }-e^{-(k+r)0})=\frac{-1}{(k+r)}(0-1)\\ =\frac{1}{(k+r)}\end{array}$(S6)

Using this method, the discounted LE will be calculated for each participant in both the standard and interventional groups.

$\begin{array}{l}\text{LE}=\underset{0}{\overset{\infty }{\int }}{e}^{-kt}dt\\ \text{LE}={|}_0^{\infty }-\frac{1}{k}{e}^{-kt}\\ \text{LE}=-\frac{1}{k}\left(e^{-k\infty }-e^{-k0}\right)\\ \text{LE}=-\frac{1}{k}\left(0-1\right)\\ \text{LE}=\frac{1}{k}\end{array}$(S7)

3) Continuous costs

The following steps are used to calculate continuous costs:

1. Convert the LE in years to LE in 30-day blocks. LE = LE_years × 365/30

2. Define the discount rate, R, in 30-day blocks. R = 0.05 × 30/365

3. Define C1, which is the mean 30-day cost for the stable phase

4. Define C2, which is the mean 30-day cost for the pre-death phase

5. Define t, which is time in 30-day blocks

6. Define x, which is time (in 30-day blocks), at which point, switch from the stable phase to the pre-death phase (note: in excel example, x = 6).

$\begin{array}{l}\text{LE}=\underset{0}{\overset{\infty }{\int }}{e}^{-kt}dt\\ \text{LE}={|}_0^{\infty }-\frac{1}{k}{e}^{-kt}\\ \text{LE}=-\frac{1}{k}\left(e^{-k\infty }-e^{-k0}\right)\\ \text{LE}=-\frac{1}{k}\left(0-1\right)\\ \text{LE}=\frac{1}{k}\end{array}$

Therefore, for any period, t, the total proportion of a patient’s LE is S(t) = e^−kt.

At time t, the proportion of patients who are in the stable phase is e^{−k(t + 6)} and the proportion of patients who are in the pre-death phase is e^−kt−e^{−k(t + 6)}.

Therefore, the total cost at any time period, t, is [C1(e^{−k(t + 6)}) + C2(e^−kt−e^{−k(t + 6)})].

To determine the cost over time period 0-∞, the following closed integral must be solved:

$\begin{array}{l}\text{Cost}={\int }_0^{\infty }\left[\text{C}1(e^{-k(t+6)})+\text{C}2(e^{-kt}-e^{-k(t+6)})\right]dt\\ ={\int }_0^{\infty }\left[\text{C}2(e^{-kt})+(C1-C2)e^{-k(t+6)}\right]dt\\ ={\int }_0^{\infty }\left[C2(e^{-kt})dt+{\int }_0^{\infty }(C1-C2)e^{-k(t+6)}\right]dt\\ =\frac{-C2}{(k)}\left(e^{-(k)\infty }-e^{-(k)0}\right)\\ +\frac{-(C1-C2)}{(k)}\left(e^{-(k)(\infty +6)}-e^{-(k)(0+6)}\right)\\ =\frac{-C2}{(k)}(0-1)+\frac{-(C1-C2)}{(k)}\left(0-e^{-6k)}\right)\\ =\frac{C2}{(k)}+\left[\frac{(C1-C2)}{(k)}\left(e^{-6k}\right)\right]\end{array}$(S8)

For discounted costs, solve the following:

$\begin{array}{l}\text{Cost}={\int }_0^{\infty }\left[\text{C}1(e^{-k(t+6)})+\text{C}2(e^{-kt}-e^{-k(t+6)})\right]*e^{-rt}dt\\ ={\int }_0^{\infty }\left[\text{C}2(e^{-kt-rt})+(C1-C2)e^{-k(t+6)-rt}\right]dt\\ ={\int }_0^{\infty }\left[\text{C}2(e^{-(k+r)t})dt+{\int }_0^{\infty }(C1-C2)e^{-k(t+6)-rt})\right]dt\\ ={\int }_0^{\infty }\left[\text{C}2(e^{-(k+r)t})dt+{\int }_0^{\infty }(C1-C2)e^{-kt-6k-rt})\right]dt\\ ={\int }_0^{\infty }\left[\text{C}2(e^{-(k+r)t})dt+{\int }_0^{\infty }(C1-C2)e^{-6k}{e}^{-kt-rt})\right]dt\\ ={\int }_0^{\infty }[\text{C}2\left[(e^{-(k+r)t})dt+{\int }_0^{\infty }(C1-C2)e^{-6k}{e}^{-(k+r)t})\right]dt\\ =\frac{-C2}{(k+r)}\left(e^{-(k+r)\infty }-e^{-(k+r)0}\right)\\ +\frac{-(C1-C2){\in }^{-\in k}}{(k+r)}\left(e^{-(k+r)\infty }-e^{-(k+r)(0)}\right)\\ =\frac{-C2}{(k+r)}(0-1)+\frac{-(C1-C2)}{(k+r)}{e}^{-6k}(0-1)\\ =\frac{C2}{(k+r)}+\left[\frac{(C1-C2)}{(k+r)}{e}^{-6k}\right]\end{array}$(S9)

Table S1 Cost categories

Table S2 Life expectancy, cumulative costs and incremental cost-effectiveness of HMI and standard care (for imputed data)

Reference

• D’AgostinoRBSrVasanRSPencinaMJGeneral cardiovascular risk profile for use in primary care: the Framingham Heart StudyCirculation2008117674375318212285
   PubMed Web of Science ®Google Scholar
• PeetersAMamunAAWillekensFBonneuxLA cardiovascular life history. A life course analysis of the original Framingham Heart Study cohortEur Heart J20022364586611863348
   PubMed Web of Science ®Google Scholar

Acknowledgments

The Hypertension Management Initiative was developed, implemented, and funded by the Heart and Stroke Foundation of Ontario, with funding support from the Ontario Ministry of Health and Long-Term Care.

Dr Krahn holds the F Norman Hughes Chair in Pharmacoeconomics at the Faculty of Pharmacy, University of Toronto. Dr Tu is supported by a Tier 1 Canada Research Chair in Health Services Research and a career investigator award from the Heart and Stroke Foundation of Ontario. This study was supported by the Toronto Health Economics and Technology Assessment Collaborative and the Institute for Clinical Evaluative Sciences, both of which are funded by an annual grant from the Ontario Ministry of Health and Long-Term Care. The opinions, results and conclusions reported in this paper are those of the authors and independent from the funding sources. No endorsement by the Institute for Clinical Evaluative Sciences or the Ontario Ministry of Health and Long-Term Care is intended or should be inferred.

Disclosure

Sheldon W Tobe reports receiving lecture and/or consulting fees from companies manufacturing antihypertensive drugs. The authors report no other conflicts of interest.