1 of 52

Cost-Effectiveness Calculators for SUD: A Pilot of Peer Recovery Support Services and Bystander Naloxone Distribution��MSD Center Webinar, October 25, 2022

Sierra Castedo de Martell, MPH, Doctoral Candidate, Sierra.J.CastedodeMartell@uth.tmc.edu

Margaret Brannon Moore, JD, LLM, MPH, Doctoral Candidate, Margaret.B.Moore@uth.tmc.edu

Hannah Wang, PhD, Programmer Analyst IV, Information Technology

H. Shelton Brown, III, PhD, Associate Professor and PI,

The University of Texas Health Science Center at Houston, School of Public Health

Funding from NIDA R24DA051988 Recovery Research Institute Pilot Grant

2 of 52

Center Resources

3 of 52

Outline for Today

Background and goals for the future
Learn about cost-effectiveness analysis
How we structured our analysis to make the calculator
Hands-on cost-effectiveness calculator tutorial – your feedback encouraged!

4 of 52

Background

Our ultimate goal:

A free, web-based multi-faceted cost-effectiveness calculator that

Empowers stakeholders (RCOs, advocates, community decision-makers) to use cost-effectiveness information
Increases support for existing programs, build support for the adoption of programs

Bonus goal:

Fill in the knowledge gaps – very little economic evaluation research on peer-driven SUD interventions

5 of 52

Background

Lots of work to do!
Unfunded collegiate recovery program calculator here
Pilot funding to make today’s calculator:

Long-term PRSS + Bystander Naloxone Distribution
Free, web-based, more accessible

Will also seek peer-reviewed publication

But reaching that goal requires a lot of work, so we’ve taken the approach of tackling just a piece of this calculator at a time. I made a collegiate recovery program cost-effectiveness calculator as part of my master’s thesis work. It’s just based in Excel, but you can take a look at it by following the link. This past year, we have had the great fortune of getting some pilot funding to make the calculator we’ll be showing you today.

That calculator estimates cost-effectiveness for long-term peer recovery support services (PRSS) and for bystander naloxone distribution, using information that you input yourself – so the result is tailored for your program the way it’s structured right now – or maybe the way you’d like a program to work. It’s also free, but it’s based on a website, and it is a lot more user-friendly and accessible compared to my Excel-based calculator.

We’ll also try to get the work we did to make the calculator published – so the underlying analysis. That will be for an academic audience.

6 of 52

Pilot funded

Background research

RCOs review model

Calculator prototype

RCOs test calculator

Finalize pilot calculator

NAPS presentation + MSDC Webinar

Your feedback

Future improvements

THANK YOU to Communities for Recovery and RecoveryATX for providing critical feedback!

Here’s a brief timeline of what we did to get here today, and where we’re going. So we got that funding, and then we did the background research to figure out how to do the cost-effectiveness analysis for PRSS. For bystander naloxone distribution, there’s actually already a cost-effectiveness study, so my colleague, Marnie, updated all those numbers. Once we figured out how we were planning to build the calculator, we met with staff at two local RCOs, Communities for Recovery and RecoveryATX to make sure we were thinking about things the way they really work. From that feedback, we built out a prototype calculator, and tested that calculator with the same two RCOs. Our IT partner, Hannah, then finalized the pilot calculator we’ll be showing you today, incorporating all that feedback. After today, we’d also love to get your feedback, and we plan to keep a feedback form open on the pilot calculator, helping us make further refinements and improvements. Eventually, we hope to add additional components, and slowly build out that multi-faceted tool we envisioned.

7 of 52

Outline for Today

Background and goals for the future
Learn about cost-effectiveness analysis
How we structured our analysis to make the calculator
Hands-on cost-effectiveness calculator tutorial – your feedback encouraged!

8 of 52

What is Cost-Effectiveness Analysis?

9 of 52

What is Cost-Effectiveness Analysis?

Resources that make an intervention happen

The intervention (program, activity)

The good stuff that our intervention produces

10 of 52

What is Cost-Effectiveness Analysis?

Resources that make an intervention happen

The intervention (program, activity)

The good stuff that our intervention produces

How balanced are resources to good stuff?

11 of 52

What is Cost-Effectiveness Analysis?

How balanced are resources to good stuff?

Cost-Effectiveness

Cost-Benefit Analysis

Return on Investment

12 of 52

What is Cost-Effectiveness Analysis?

How balanced are resources to good stuff?

Cost-Effectiveness

Cost-Benefit Analysis

Return on Investment

Both resources and good stuff have $$$

Resources have $$$, but good stuff doesn’t

13 of 52

What is Cost-Effectiveness Analysis?

How balanced are resources to good stuff?

Cost-Effectiveness

Cost-Benefit Analysis

Return on Investment

Both resources and good stuff have $$$

Resources have $$$, but good stuff doesn’t

14 of 52

15 of 52

Grocery store metaphor:

Compare sticker prices, but packaging or product is not identical, so we can compare price per ounce (or other unit), instead.

16 of 52

Grocery store metaphor:

Compare sticker prices, but packaging or product is not identical, so we can compare price per ounce (or other unit), instead.
For different types of cereal.

17 of 52

Grocery store metaphor:

Compare sticker prices, but packaging or product is not identical, so we can compare price per ounce (or other unit), instead.
For different types of cereal.

18 of 52

Grocery store metaphor:

Compare sticker prices, but packaging or product is not identical, so we can compare price per ounce (or other unit), instead.
Or for the exact same product and brand, but different sizes (economies of scale)

19 of 52

20 of 52

21 of 52

22 of 52

4 years perfect health

QOL weight = 1

4 x 1 = 4

= 4 QALYs added

4 years at half of perfect health

QOL weight = 0.5

4 x 0.5 = 2

= 2 QALYs added

23 of 52

24 of 52

25 of 52

26 of 52

27 of 52

28 of 52

29 of 52

30 of 52

31 of 52

32 of 52

33 of 52

34 of 52

35 of 52

Dealing with Uncertainty

Base Case

Multi-Way Sensitivity Analysis

One-Way Sensitivity Analysis

36 of 52

Dealing with Uncertainty

Base Case: Our basic model for a set time period. We’re not looking at any uncertainty here, we’re just using whatever numbers we have, usually an average or a median.

Base Case

Multi-Way Sensitivity Analysis

One-Way Sensitivity Analysis

Calculator

37 of 52

Dealing with Uncertainty

One-Way Sensitivity Analysis: Change one input at a time: how does cost-effectiveness change when input changed (for example: more participants, higher cost of naloxone, better retention of participants)

Base Case

Multi-Way Sensitivity Analysis

One-Way Sensitivity Analysis

Calculator

38 of 52

Dealing with Uncertainty

Base Case

Multi-Way Sensitivity Analysis

One-Way Sensitivity Analysis

Calculator

Full evaluation or academic papers

39 of 52

Outline for Today

Background and goals for the future
Learn about cost-effectiveness analysis
How we structured our analysis to make the calculator
Hands-on cost-effectiveness calculator tutorial – your feedback encouraged!

40 of 52

Moving into the Models

Bystander naloxone distribution - Marnie

Long-term, post-treatment PRSS delivered in an RCO setting - Sierra

41 of 52

Bystander Naloxone Distribution

People who use opioids

No bystander Naloxone kit present

% who do not overdose

% who overdose and survive

% who overdose and die

Bystander Naloxone kit present

% who do not overdose

% who overdose and survive

% who overdose and die

Updated from Coffin and Sullivan, 2013

42 of 52

Bystander Naloxone Distribution

Cost of Naloxone Kit

Medical and associated costs for OD with bystander naloxone

Medical and associated costs for OD with no bystander naloxone

QALE under BND

QALE under tx only

tx = usual treatment (e.g., EMS, ED treatment)

QALE = quality-adjusted life expectancy

HEALTH SYSTEM

When calculating the cost-effectiveness of a bystander naloxone program, there are a number of parameters to consider. Reducing those parameters to a simple equation results in the equation shown on this slide. The numerator is the cost of the naloxone kit plus the medical costs associated with naloxone administration minus the medical costs if no naloxone is administered by a bystander. The denominator is the quality adjusted life expectancy with bystander naloxone in place minus the quality-adjusted life expectancy if there is no bystander naloxone. There often may be naloxone costs on both sides of the numerator, as those who don’t received it from a bystander may receive it from a provider, either a paramedic or doctor once transported to the hospital. Part of the difference in the numerator stems from the increase in transport refusals from those who have been revived by naloxone. The primary difference in the equation, though, is in the denominator – the number of lives saved due to the bystander naloxone.

43 of 52

Bystander Naloxone Distribution

Recall: Health system perspective versus societal perspective

Health system – just the costs that would be carried by whatever the health system is. Care about averted medical costs within the health system, too.
Societal – Need to think about things like someone’s time outside of the health system, productivity, etc.

With bystander naloxone distribution programs, a significant factor for the societal perspective is lives saved.

44 of 52

Long-Term PRSS

People who get specialty SUD treatment

Not getting PRSS after treatment

% who stay in recovery

% who return to chaotic use

% who die

Get 1 year of PRSS after treatment

Same as above + drop out (re-enter normal risk pool)

And let’s talk about PRSS

If you really think about it, standard care for SUD is nothing at all. Any given year 95% get nothing, not a doctor’s visit, not a chat with a friend – nothing. Lots of reasons for that, including not being interested or ready, but let’s set that aside.
So, it’s probably not fair to set up our analysis as PRSS versus nothing. PRSS would look way too good, by comparison. So we’ll do the next best thing – specialty SUD treatment. We know most people don’t get it, but it is a really good option and about as good as we can do right now on any kind of large scale. So in our analysis, we include both inpatient and outpatient treatment, we’re kind of averaging across both kinds of treatment.

Picture it as someone enters specialty treatment, and when they finish, they can either get long-term PRSS or not.

We set this up as 1 year, but the calculator is asking you for 1 year or completion/graduation – think about it that way
At the end of that year, some % will still be in recovery (so, reduced or non-chaotic use, or abstinence, as defined by the org inputting these numbers). There’s some % who returns to chaotic use. And there is a smaller % that dies. That applies to both groups. So all of our costs and our estimates of the good stuff are based on what’s happening with these three groups.

45 of 52

Long-Term PRSS

People who get specialty SUD treatment

Not getting PRSS after treatment

% who stay in recovery

% who return to chaotic use

% who die

Get 1 year of PRSS after treatment

Same as above + drop out (re-enter normal risk pool)

Treatment as Usual (TAU)

Intervention

46 of 52

Long-Term PRSS

Cost of PRSS for 1 year after treatment

Return to active SUD under PRSS * cost of tx * 10%

Averted medical costs, PRSS

Return to active SUD under tx only * cost of tx * 10%

Averted medical costs, tx only

Effect under PRSS

Effect under tx only

tx = specialty SUD treatment

HEALTH SYSTEM

47 of 52

Long-Term PRSS

Cost of PRSS for 1 year after treatment

Return to active SUD under PRSS * cost of tx * 10%

Averted medical costs, PRSS

Return to active SUD under tx only * cost of tx * 10%

Averted medical costs, tx only

Effect under PRSS

Effect under tx only

tx = specialty SUD treatment

QALYs

People in recovery at 3 years

HEALTH SYSTEM

For the bottom part, we looked at both quality adjusted life years – adding up added years of life across those three groups – recovery, chaotic use, and death, and then adjusted for the differences in quality of life across those groups. And we also looked at number of people in recovery after 3 years. So for that we just added up people in the recovery group for PRSS and for TAU.

We added up the costs of PRSS episodes – again an average picture, taking into account all the variation in how many hours, what people are paid or reimbursed, etc.
And we also added the cost of maybe having to go back to specialty treatment for SUD. So that’d be just those who returned to chaotic use in spite of PRSS. And then we also know that not everyone’s going to treatment again, so we accounted for that – assumed that only about 10% of people in chaotic use get specialty tx.
And then we subtracted out averted medical costs under PRSS.
Then we subtracted the treatment as usual costs. We didn’t include the cost of that initial treatment round, because remember, we’re modeling this as though everyone got treatment – treatment is our starting point, so it’d be the same on both sides, and zero out. Instead we start with kind of the complementary thing to this block – who maybe got another round of treatment from the TAU group? Again assuming about 10% get tx, and only among those who returned to chaotic use. We also do have averted medical costs under TAU, so we include that too.

48 of 52

Long-Term PRSS

Cost of PRSS for 1 year after treatment, + participant time

Return to active SUD under PRSS * cost of tx *10%

Averted societal costs (not medical, not productivity), PRSS

Return to active SUD under tx only * cost of tx *10%

Averted societal costs (not medical, not productivity), tx only

Effect under PRSS

Effect under treatment only

tx = specialty SUD treatment

QALE = quality-adjusted life expectancy

QALYs

People in recovery at 3 years

SOCIETAL

49 of 52

Outline for Today

Background and goals for the future
Learn about cost-effectiveness analysis
How we structured our analysis to make the calculator
Hands-on cost-effectiveness calculator tutorial – your feedback encouraged!

50 of 52

Let’s look at the calculator!

web.sph.uth.edu/cea/

51 of 52

Additional feedback or questions?

H.Shelton.Brown@uth.tmc.edu

Sierra.J.CastedodeMartell@uth.tmc.edu

Margaret.B.Moore@uth.tmc.edu

Please take our feedback survey!

https://redcap.link/calculator

52 of 52

Download these slides

https://bit.ly/CEACalculatorSlides