Cost-Effective Is Not Affordable, and Efficient Is Not Fair: Budget Impact and Equity

อ่านฉบับภาษาไทย (Thai version)

Abstract

A district hospital committee has found a new preventive treatment, C, good value: 150,000 THB per extra quality-adjusted life year (QALY), under the 160,000 THB the committee will pay for one extra QALY, its threshold. This article asks three questions that number cannot answer. Can the payer afford it? A budget impact analysis turns 10,000 eligible people, a rising share on C and 50,000 THB extra per person-year into a yearly bill: 100, 200 and 300 million THB against a yearly budget of 200 million. What does the money push out? If the care it replaces buys a QALY for 160,000 THB, the 600 million that buys 4,000 QALYs on C removes 3,750 elsewhere, a net gain of 250 QALYs. Who gains? Starting C near the hospital gives the town 925 QALYs and costs rural districts 675; an equal rollout gives each 25. Equity weights, severity, hardship and the limits of the QALY show what the committee must weigh in words.

Summary card: the title, the key idea that good value, affordability and fairness are three separate questions, and five takeaways for treatment C in one province: 150,000 THB per extra QALY, yet a yearly bill of 100, 200 and 300 million THB against a budget of 200 million; 10,000 eligible × 60% on C × (60,000 − 10,000) THB = 300 million THB in year 3, with the saving counted once; if the care pushed out buys a QALY for 160,000 THB, 600 million THB buys 4,000 QALYs on C but pushes out 3,750 elsewhere, a net gain of 250 QALYs, and town first gives the town +925 QALYs and rural districts −675 QALYs; the equal plan wins once a rural QALY counts more than 1.286 times; the same extra year counts 0.80 QALYs in good health and 0.40 with a disability. — Visual summary · ภาพสรุป

Good value, but can the fund pay?

A district hospital committee has decided that C, a new preventive treatment for adults with a chronic condition, is good value. Per patient, C costs 150,000 THB (Thai baht) more than usual care and adds 1.00 QALY. A QALY, a quality-adjusted life year, is a year of life weighted by its quality, from 1 for full health to 0 for death QALY, DALY and the Utilities Behind Them: What Is Actually Being Measured.

So each extra QALY costs 150,000 THB. That price per extra QALY is the incremental cost-effectiveness ratio (ICER). The most the committee is willing to pay for one extra QALY is 160,000 THB, a benchmark called the threshold, so C passes; the full lesson is in the ICER article.

The committee's verdict now goes to the payer, the health fund that pays for care across the province, where 10,000 adults meet C's criteria. C is a course of three years. Each year on it costs the fund 60,000 THB for the drug, and fewer complications save 10,000 THB of hospital care in the same year.

So the fund pays 50,000 THB more per person for each year on C, and 150,000 THB over the course: the same 150,000 THB as in the ICER. The fund has set aside 200 million THB a year for new treatments.

The fund manager asks three questions that the ICER cannot answer. How much will C cost the fund in each of the next three years? What care will the fund stop paying for to make room? And who gets C first: people in the main town, near the provincial hospital, or people in the rural districts?

A common shortcut reads "under 160,000 THB per QALY" as "pay for it for everyone". But the ICER is a price per patient. It does not know how many patients there are, when the bills arrive, or who receives the treatment. So this article answers the three questions with the province's numbers.

From one patient to a yearly bill

Read first on this site:

Start with how many people are on C in each year. The province's clinics can start 2,000 people a year, and each stays on C for three years. So 2,000 are on C in year 1, 4,000 in year 2 and 6,000 in year 3: 20%, 40% and 60% of the 10,000 eligible. The share of eligible people who actually receive a treatment is called its uptake.

Each person-year on C, one person treated for one year, costs the fund 50,000 THB more than usual care. So the fund pays 2,000 × 50,000 = 100 million THB in year 1, 200 million in year 2 and 300 million in year 3: 600 million THB over three years. Year 3 needs 100 million THB more than the fund's budget allows.

Behind each year's bill are two amounts. In year 3 the drug alone costs 6,000 × 60,000 = 360 million THB, and 6,000 × 10,000 = 60 million THB of complication care is saved, so the fund pays 360 − 60 = 300 million THB. C is added to usual care and replaces no other drug, so nothing else is subtracted. A drug it replaced would be subtracted the same way as the saving.

This yearly bill has a name. A budget impact analysis estimates how a payer's spending changes, year by year, when it adopts a new treatment [1, 2]. It takes the payer perspective: it counts only the money this fund pays, not patients' travel or lost work. It usually covers one to five years, because budgets are set year by year.

For each year t, with B_t the fund's bill in year t, N the eligible people, s_t the uptake in year t, d the drug cost per person-year and o the care it saves per person-year:

$$B_t = N \times s_t \times (d - o)$$

In year 3 that is 10,000 × 60% × (60,000 − 10,000) = 300 million THB.

⤢ click to enlarge

Look at the black marks: they are what the fund pays each year, 100, 200 and 300 million THB. Blue bars above zero are C's drug and green bars below zero are the care it saves, so each black mark is blue minus green. In year 3, blue 360 minus green 60 gives the black mark at 300 million THB. Only year 3 rises above the dashed budget line, by 100 million THB.

Count a saving once

The 10,000 THB of complication care that C saves each year is a cost offset: money the payer no longer spends elsewhere because of the new treatment. It is already inside the 50,000 THB, which is the drug minus the saving. Subtracting year 3's 60 million THB a second time, as a separate saving, would leave 300 − 60 = 240 million THB and hide most of the overrun.

Every year at full value

Lifetime models, such as the one behind the ICER, often shrink amounts that arrive later. Discounting is converting a future cost or health gain into its value today, shrinking it by the discount rate for each year away. At 3% a cost of 1,834 THB due in year 3 is worth 1,678 THB today; the full lesson is in the costing article.

A budget plan does not do this. In year 3 the fund pays the full 300 million THB in cash, so every year stays at full value. This article assumes that the ICER's 150,000 THB is the plain sum 3 × 50,000, with no shrinking, so that the bill and the ICER can be set side by side. A discounted ICER would need its own sums.

The eligible count is held at 10,000 here: each year as many people are newly diagnosed as die, recover or move away, so the pool neither fills nor drains. A real plan counts the two flows year by year, as the counting lesson linked above does [1].

R code yearly_bill.R

# The fund's yearly bill for C: 2,000 people start each year and stay on C for three years.
eligible <- 10000
on_c <- 2000 * 1:3                    # people on C in years 1, 2 and 3
drug <- 60000; offset <- 10000        # THB per person-year: C's drug; complication care it saves
extra <- drug - offset
bill <- on_c * extra / 1e6            # million THB
cat("People on C:", on_c, "| share of the eligible (%):", 100 * on_c / eligible, "\n")
cat("Drug (million THB):", on_c * drug / 1e6, "\n")
cat("Offsets (million THB):", -on_c * offset / 1e6, "\n")
cat("Net bill (million THB):", bill, "| sum", sum(bill), "| year 3 over 200:", bill[3] - 200, "\n")
cat("Extra per person:", extra, "THB a year,", 3 * extra, "THB over the three years\n")
cat("1,834 THB due in year 3, valued today at 3% a year:", round(1834 / (1 + 0.03)^3), "THB\n")

Output of the run yearly_bill.log

People on C: 2000 4000 6000 | share of the eligible (%): 20 40 60
Drug (million THB): 120 240 360
Offsets (million THB): -20 -40 -60
Net bill (million THB): 100 200 300 | sum 600 | year 3 over 200: 100
Extra per person: 50000 THB a year, 150000 THB over the three years
1,834 THB due in year 3, valued today at 3% a year: 1678 THB

The fund's bill for three years: drug, offsets, net bill and the year-3 overrun, then the discounting example from the recap.

What the money would have bought

The fund's budget is fixed, so the 600 million THB for C has to come from care the fund would otherwise have paid for, such as clinic visits, screening or other medicines: the care that C pushes out. That care also produced health. The health lost when money moves away from other care is called the health opportunity cost [3, 4]. Here all 600 million THB counts as money that would otherwise have paid for such care, whether it comes from the new-treatment budget or from cuts elsewhere.

Suppose the care pushed out buys one QALY for every 160,000 THB, the same as the committee's threshold. Then 600 million THB of that care would have bought 600 million ÷ 160,000 = 3,750 QALYs. Spent on C at 150,000 THB per QALY, the same money buys 600 million ÷ 150,000 = 4,000 QALYs. Each person-year on C, at 50,000 THB, buys a third of the course's 1.00 QALY, if that QALY is counted evenly over the three years, an assumption.

The province gains 4,000 − 3,750 = 250 QALYs overall, far fewer than the 4,000 QALYs that C's patients gain. This difference is the net health benefit (NHB): the health a treatment adds minus the health lost in the care it pushes out [4]. With S the money spent and k the cost per QALY of the care pushed out:

$$NHB=\frac{S}{ICER}-\frac{S}{k}$$

It is positive only when C's cost per QALY is below k. The threshold and k are different numbers. The threshold is what the committee chooses to pay for a QALY; k is what the care pushed out really costs per QALY, a fact of the health system that nobody knows exactly. If the care pushed out buys a QALY for 120,000 THB, the same money would have bought 5,000 QALYs there, and C costs the province 1,000 QALYs net.

At k = 150,000 THB, C's own cost per QALY, gain and loss cancel and the net is zero. So "under the threshold" means "more health overall" only if the threshold is close to what the care pushed out really costs per QALY.

R code net_health.R

# What 600 million THB buys on C, and what it would have bought in the care it pushes out.
spend <- 600e6                        # THB over the three years, from the yearly bill
icer <- 150000                        # THB per extra QALY on C
gained <- spend / icer
cat("QALYs gained on C:", gained, "\n")
for (k in c(160000, 120000, 150000))  # THB per QALY in the care that is pushed out
  cat("Care pushed out at", k, "THB per QALY: lost", spend / k, "| net", gained - spend / k, "\n")

Output of the run net_health.log

QALYs gained on C: 4000
Care pushed out at 160000 THB per QALY: lost 3750 | net 250
Care pushed out at 120000 THB per QALY: lost 5000 | net -1000
Care pushed out at 150000 THB per QALY: lost 4000 | net 0

What 600 million THB buys on C against what it held in the care pushed out, at three costs per QALY.

The controls start at the province: 10,000 eligible people, 60% on C by year 3, 50,000 THB extra a year and a budget of 200 million THB. Look at the left panel: only year 3's bar rises above the budget line. Drag the cost per QALY of the care pushed out down to 120,000 THB and watch the net bar on the right fall below zero, because that money would have bought more health elsewhere than it buys on C.

Price, the bill and the Thai committees

Read first on this site:

How economic evaluation feeds national drug selection in Thailand

Thai guidance for health technology assessment asks for a budget impact analysis beside the economic evaluation [5]. The committees that decide the national list of essential medicines and the benefit package of universal coverage weigh value, the yearly bill and whether the health system can deliver the treatment.

Price is the lever both questions share. If C's drug falls from 60,000 to 40,000 THB a year, the fund pays 30,000 THB extra per person-year. The bill becomes 60, 120 and 180 million THB, under the budget every year. Over three years it is 360 million THB.

C's cost per QALY falls to 3 × 30,000 = 90,000 THB, and 360 million ÷ 90,000 = 4,000 QALYs, the same health as before. The care pushed out shrinks to 360 million ÷ 160,000 = 2,250 QALYs, so the net health benefit rises from 250 to 1,750 QALYs.

Slower uptake also lowers the bill, but it delays the health and, as the next section shows, decides who waits. And a lower price does not train nurses or open clinics: whether the system can deliver C is a separate check.

R code price.R

# Drug at 40,000 THB a year instead of 60,000; the complication care saved stays 10,000 THB.
on_c <- 2000 * 1:3                    # people on C in years 1, 2 and 3
extra <- 40000 - 10000                # THB extra per person-year
bill <- on_c * extra / 1e6            # million THB
cat("Drug at 40,000 THB:", extra, "THB extra a year | net bill (million THB):", bill,
    "| sum", sum(bill), "\n")
spend <- sum(bill) * 1e6              # THB over the three years
icer <- 3 * extra / 1.00              # THB per extra QALY: three years of extra cost for 1.00 QALY
cat("Cost per QALY:", icer, "| gained", spend / icer, "| lost at 160,000 THB per QALY",
    spend / 160000, "| net", spend / icer - spend / 160000, "\n")

Output of the run price.log

Drug at 40,000 THB: 30000 THB extra a year | net bill (million THB): 60 120 180 | sum 360
Cost per QALY: 90000 | gained 4000 | lost at 160,000 THB per QALY 2250 | net 1750

Drug at 40,000 THB. Look at the last line: the same 4,000 QALYs gained now sit against 2,250 lost, a net of 1,750.

Who gets C first?

Half of the 10,000 eligible people live in the main town, around the provincial hospital, and half in rural districts. The clinic that starts C is in that hospital, and rural patients must travel to reach it. So in the first plan the clinic starts 1,400 town patients and 600 rural patients a year. The steps below follow this town-first plan and keep the earlier supposition that the care pushed out buys a QALY for 160,000 THB, which gives 3,750 QALYs lost.

The town-first plan, step by step

Each year's starters stay on C for three years, so within the three years a year-1 starter counts 3 person-years, a year-2 starter 2 and a year-3 starter 1.

Person-years on C

\[ 1400\times(1+2+3)=8400,\ 600\times(1+2+3)=3600 \]

Town, then rural districts.
QALYs gained

\[ 8400\div 3=2800,\ 3600\div 3=1200 \]

Each person-year on C buys a third of a QALY.
QALYs lost

\[ 3750\div 2=1875 \]

The care pushed out is shared equally by the two groups, an assumption.
Net, by group

\[ 2800-1875=925,\ 1200-1875=-675 \]

What each group ends with.

Result: The town ends 925 QALYs better off and rural districts 675 QALYs worse off: +250 QALYs in total, but the gap between them, the town's net minus the rural net, widens by 925 − (−675) = 1,600 QALYs.

What the total hides

Counting gains and losses group by group like this, not only in total, is called distributional cost-effectiveness analysis [3]. Here it shows what the total of +250 QALYs hides: rural districts lose more care than they gain from C.

A second plan starts 1,000 people a year in each group and pays 32 million THB for a mobile clinic to reach rural patients. Each group has 1,000 × 6 = 6,000 person-years and gains 6,000 ÷ 3 = 2,000 QALYs. The fund now spends 600 + 32 = 632 million THB, which would have bought 632 million ÷ 160,000 = 3,950 QALYs elsewhere, 1,975 per group. So each group ends 2,000 − 1,975 = 25 QALYs better off: 50 QALYs in total, with the gap unchanged.

So the choice is not between a good plan and a bad one. The town-first plan gives 250 − 50 = 200 more QALYs in total, while the equal plan keeps the gap from widening by 1,600 QALYs. Drawing both plans on one picture, total health across and the change in the gap up, gives the equity impact plane [3].

⤢ click to enlarge

Read across for total health and up for a narrower gap between town and rural. The town-first plan sits lower right: more health in total but a wider gap, because the town gains 925 QALYs and rural districts lose 675 QALYs. Follow the arrow to the equal plan: it gives up 200 QALYs in total so that the gap does not widen.

R code who_gains.R

# Two ways to roll C out: starts per year in town and in rural districts (5,000 eligible in each).
icer <- 150000; k <- 160000; extra <- 50000; cut <- 13 / 60 - 8 / 66   # cut: the trial's risk gap
starts <- cbind(town_first = c(1400, 600), equal = c(1000, 1000))     # rows: town, rural
years <- starts * (1 + 2 + 3); gain <- years * extra / icer   # person-years on C; QALYs gained
spend <- colSums(years) * extra + c(0, 32e6)                  # THB; 32 million: the mobile clinic
net <- gain - rep(spend / k / 2, each = 2)                    # care pushed out, shared equally
cat("Person-years, town and rural, town first then equal:", years, "| gained:", gain, "\n")
cat("Spend (million THB):", spend / 1e6, "| lost:", spend / k, "| each group:", spend / k / 2, "\n")
cat("Net:", net, "| total:", colSums(net), "| gap change:", net[1, ] - net[2, ], "\n")
tie <- c(net[1, 1] - net[1, 2], net[2, 2] - net[2, 1])   # switching: town gives up; rural gains
cat("Rural weight at a tie:", tie[1], "/", tie[2], "=", round(tie[1] / tie[2], 3), "\n")
cat("One-year risk per 100 patients:", round(100 * c(13 / 60, 8 / 66, cut), 2), "| rural avoided:",
    round(years[2, ] * cut), "| families spared hardship:", round(years[2, ] * cut / 4), "\n")

Output of the run who_gains.log

Person-years, town and rural, town first then equal: 8400 3600 6000 6000 | gained: 2800 1200 2000 2000
Spend (million THB): 600 632 | lost: 3750 3950 | each group: 1875 1975
Net: 925 -675 25 25 | total: 250 50 | gap change: 1600 0
Rural weight at a tie: 900 / 700 = 1.286
One-year risk per 100 patients: 21.67 12.12 9.55 | rural avoided: 344 573 | families spared hardship: 86 143

Two rollouts, town and rural: person-years, QALYs gained, money spent and QALYs lost, the net, the gap, the rural weight at a tie, the trial's risks and the complications avoided.

How much should a rural QALY count?

One way to choose is to ask how much a QALY for the worse-off group should count. With every QALY counted the same, the town-first plan wins, 250 to 50.

Now let each rural QALY count w times. The town-first plan's weighted total is the town's gain plus w times the rural loss, 925 − 675w, and the equal plan's is 25 + 25w. They tie when 925 − 675w = 25 + 25w, that is when 900 = 700w: switching to the equal plan costs the town 925 − 25 = 900 QALYs and gives rural districts 25 + 675 = 700 QALYs. So w = 900 ÷ 700 = 1.286.

A multiplier like w on one group's QALYs is called an equity weight. A committee that counts a rural QALY as worth more than 1.286 town QALYs should prefer the equal plan. The weight is a value judgment, not a fact the data can supply, so report the unweighted result, the weighted one and the weight that flips the choice [3].

Severity, rescue and rare diseases

Weights are also argued for severity: how much healthy life a condition takes away. A person of the same age without the condition can expect about 12.50 QALYs (an illustrative figure), while a patient on usual care can expect 5.00, the usual-care figure behind the ICER. The condition takes 7.50 of the 12.50, a share of 60% called the proportional shortfall. Some payers give extra weight to conditions with a large shortfall [6].

Now picture one identifiable patient, a child with a rare disease, whose drug would cost 1,600,000 THB for each QALY it adds: 10 times the committee's threshold of 160,000 THB. Few committees find it easy to say no to a named child. That strong pull to save an identifiable person facing death, whatever the cost per QALY, is called the rule of rescue.

Rare diseases raise it often, with few patients, small trials and high prices. Rarity alone does not make a condition severe or a treatment effective, so a payer that gives such cases extra weight should say so and say why.

Protection from hardship

Health is not the only thing C protects. A complication means a hospital stay, and for a rural family the travel, lost work and extra payments can take a large part of its income. For a family earning 240,000 THB a year, a complication that costs it more than 24,000 THB takes over a tenth of that income. Health spending above a set share of a household's income, often a tenth, is called catastrophic expenditure, and keeping families below that line is financial risk protection.

In the trial behind C, 13 of 60 patients on usual care had a complication within a year, 21.67 per 100 patients, against 8 of 66 on C, 12.12 per 100 patients: 9.55 fewer per 100 people treated for a year. These are risks over one year, shares of patients each counted once, not rates per unit of time. If C cuts that one-year risk by the same 9.55 points in every year of the course, and each person-year counts at most one complication, the town-first plan's 3,600 rural person-years avoid 3,600 × 9.55 ÷ 100 = 344 complications. The equal plan's 6,000 rural person-years avoid 6,000 × 9.55 ÷ 100 = 573 complications.

Suppose one rural family in four that faces a complication passes the line. Then the plans spare about 344 ÷ 4 = 86 and 573 ÷ 4 = 143 families. Reporting health and money protected group by group, often by income, is called extended cost-effectiveness analysis [3]. The one-in-four share here is an assumption for teaching; a real estimate needs household data on payments and income.

What a QALY leaves out

Every number so far counts QALYs. A QALY values each year by the quality it is lived in, and it counts every QALY the same, whoever gains it. That is its strength and its limit [6].

Take two patients who each gain one more year of life from C. For one the year is lived in good health for this age, valued 0.80; the other also lives with an unrelated disability, and the year is valued 0.40. The QALY count gives the first 0.80 QALYs and the second 0.40, so treating the second looks half as worthwhile, although both gained the same year.

A health-state value measures how a state is lived, not what a person is worth. Age works the same way: a person with 30 years ahead can gain more QALYs than one with 5. So cost per QALY tends to favour younger patients. QALYs also leave out what falls outside the patient's own health: carers' time, a family's hardship, and the comfort of care near home.

None of this makes the ICER useless. It means the committee should see the value, the yearly bill, the net health and the split by group side by side, and say in words how it weighed them [6].

R code qaly_limits.R

# Severity, rescue, the hardship line and the capacity to gain health.
healthy <- 12.50; usual <- 5.00   # QALYs left at one age: without the condition; on usual care
shortfall <- healthy - usual
cat("Shortfall:", formatC(shortfall, format = "f", digits = 2), "QALYs, or",
    100 * shortfall / healthy, "% of", formatC(healthy, format = "f", digits = 2), "\n")
rescue <- 1600000                 # THB per QALY of a drug for one identifiable patient
cat("Rescue drug against the 160,000 THB threshold:", rescue / 160000, "times\n")
income <- 240000                  # THB a year, one rural family
cat("Hardship line at a tenth of income:", format(income / 10, big.mark = ","), "THB\n")
well <- 0.80; disabled <- 0.40    # one extra year of life: in good health; with a disability
cat("The same extra year counts:", formatC(c(well, disabled), format = "f", digits = 2), "QALYs\n")

Output of the run qaly_limits.log

Shortfall: 7.50 QALYs, or 60 % of 12.50
Rescue drug against the 160,000 THB threshold: 10 times
Hardship line at a tenth of income: 24,000 THB
The same extra year counts: 0.80 0.40 QALYs

Severity as a shortfall of QALYs, the rescue drug against the threshold, the hardship line, and one extra year valued for two patients.

Three ways to misread a good ICER

Reading good value as affordable

C costs 150,000 THB per extra QALY, yet its year-3 bill is 300 million THB against a budget of 200 million THB.

Fix: Show the yearly bill for one to five years beside the ICER.
Counting a saving twice

The 10,000 THB of care saved is already inside the 50,000 THB extra; subtracting it again shrinks year 3 to 240 million THB.

Fix: Subtract only savings that the cost difference does not already contain.
Reading the total as fair

If the care pushed out buys a QALY for 160,000 THB, both plans add health in total, but the town-first plan leaves rural districts 675 QALYs worse off.

Fix: Report gains and losses by group, and the weight that would flip the choice.

Terms as we use them

Budget impact analysis (การวิเคราะห์ผลกระทบด้านงบประมาณ): The change in a payer's spending, year by year, when it adopts a treatment; here 100, 200 and 300 million THB.
Payer perspective (มุมมองของผู้จ่าย): Counting only the money one payer spends, not patients' own costs.
Uptake (สัดส่วนการใช้): The share of eligible people who actually receive the treatment; 20%, 40% and 60% here.
Cost offset (เงินที่ประหยัดได้): Money the payer no longer spends elsewhere because of the treatment, counted once.
Health opportunity cost (ค่าเสียโอกาสทางสุขภาพ): The health lost in the care that a new spending pushes out.
Net health benefit (ประโยชน์สุขภาพสุทธิ): The health a treatment adds minus the health lost in the care it pushes out.
Distributional cost-effectiveness analysis (การวิเคราะห์ต้นทุนประสิทธิผลเชิงการกระจาย): Counting health gained and lost group by group, not only in total.
Equity impact plane (ระนาบผลกระทบด้านความเป็นธรรม): A picture with total health across and the change in a health gap up.
Equity weight (น้ำหนักความเป็นธรรม): A multiplier on one group's QALYs; a value judgment, not a measurement.
Proportional shortfall (สัดส่วนสุขภาพที่ขาดหาย): The share of the healthy life a person could expect that a condition takes away.
Rule of rescue (หลักการช่วยชีวิต): The pull to save an identifiable person facing death, whatever the cost.
Catastrophic expenditure (รายจ่ายด้านสุขภาพแบบหายนะ): Health spending above a set share of a household's income, often a tenth.
Financial risk protection (การคุ้มครองความเสี่ยงทางการเงิน): Keeping households below that line when illness strikes.
Extended cost-effectiveness analysis (การวิเคราะห์ต้นทุนประสิทธิผลแบบขยาย): Reporting health and money protected group by group, often by income.

References

Sullivan SD, Mauskopf JA, Augustovski F, Jaime Caro J, Lee KM, Minchin M, Orlewska E, Penna P, Rodriguez Barrios JM, Shau WY. Budget impact analysis-principles of good practice: report of the ISPOR 2012 Budget Impact Analysis Good Practice II Task Force. Value Health. 2014;17(1):5-14. doi:10.1016/j.jval.2013.08.2291
Mauskopf JA, Sullivan SD, Annemans L, Caro J, Mullins CD, Nuijten M, Orlewska E, Watkins J, Trueman P. Principles of good practice for budget impact analysis: report of the ISPOR Task Force on good research practices-budget impact analysis. Value Health. 2007;10(5):336-347. doi:10.1111/j.1524-4733.2007.00187.x
Cookson R, Mirelman AJ, Griffin S, Asaria M, Dawkins B, Norheim OF, Verguet S, Culyer AJ. Using cost-effectiveness analysis to address health equity concerns. Value Health. 2017;20(2):206-212. doi:10.1016/j.jval.2016.11.027
Stinnett AA, Mullahy J. Net health benefits: a new framework for the analysis of uncertainty in cost-effectiveness analysis. Med Decis Making. 1998;18(2 Suppl):S68-S80. doi:10.1177/0272989X98018002S09
Chaikledkaew U, Kittrongsiri K. Guidelines for health technology assessment in Thailand (second edition) - the development process. J Med Assoc Thai. 2014;97 Suppl 5:S4-S9. https://pubmed.ncbi.nlm.nih.gov/24964693/
Drummond MF, Sculpher MJ, Claxton K, Stoddart GL, Torrance GW. Methods for the Economic Evaluation of Health Care Programmes. 4th ed. Oxford: Oxford University Press; 2015. ISBN 9780199665884

The series, in learning order

Key takeaways

The ICER is a price per patient, 150,000 THB per extra QALY; the payer's yearly bill is 100, 200 and 300 million THB against a budget of 200 million a year.
Build the bill from people, uptake and the extra cost: 10,000 × 60% × (60,000 − 10,000) = 300 million THB in year 3, with the saving counted once.
Money spent on C leaves other care: at 160,000 THB per QALY, the 600 million THB that buys 4,000 QALYs on C pushes out 3,750, a net gain of only 250 QALYs.
Totals hide who gains: with the care pushed out at 160,000 THB per QALY, town first gives the town +925 QALYs and rural districts −675 QALYs; the equal plan wins once a rural QALY counts more than 1.286 times.
A QALY values how a year is lived, not who lives it: the same extra year counts 0.80 QALYs in good health and 0.40 with a disability.