Table 1 Summary of simulation model parameters

β₁

Absolute increase in treatment effect on cure when X₁ = 1

.024

β₂

Absolute increase in treatment effect on cure when X₂ = 1

.048

β₃

Absolute increase in treatment effect on cure when X₃ = 1

.071

β₄

Absolute increase in treatment effect on cure when X₄ = 1

.095

β₅

Absolute increase in treatment effect on cure when X₅ = 1

.119

β₆

Absolute increase in treatment effect on cure when X₆ = 1

.143

TE_i

True treatment effect on outcome for patient “i” as a function of X_1i,X_2i,X_3i,X_4i,X_5i,X_6i

Ranges from 0 to .5. Distribution in Fig. 2

S_i

Patient “i” severity level directly effecting cure but have no effect on treatment effectiveness and are unrelated to treatment choice

Distributed Uniform(-.5,.5)

α₀

Untreated patient cure probability at mean severity level

.1

α_S

Change in untreated patient cure probability given S_i

-.1

V

The value patients gain when cured

$800

C

The cost of treatment

$200

K_j

The proportion of knowledge of treatment effectiveness that is patient-specific in simulation “j”

\(\left(\begin{array}{c}{0}\text{,} \, \text{.}{1}\text{,} \, \text{.}{2}\text{,} \, \text{.}{3}\text{,} \, \text{.}{4}\text{,} \, \text{.}{5}\text{,} \, \text{.}{6}\text{,} \, \\ \text{.}{7}\text{,} \, \text{.}{8}\text{,} \, \text{.}{9}\text{,} \, {1}\end{array}\right)\)

U_i

Accumulated unmeasured factors for patient “i” which affect treatment valuation

N(0,25)

ETE_i

Expected treatment effect for patient “i” given knowledge within simulation “j”

K_j * (TE_i—.25) + .25

EVT_i

Expected value of treatment for patient “i” given ETE_i

V• ETE_i + C + U_i

T_i

1 if patient is EVE_i is greater than 1, 0 otherwise

P(Y_i|T_i,S_i)

Probability patient “i” is cured given TE_i, T_i, and S_i

.1 + TE_i•T_i + (-.1)•S_i

Y_i

1 if patient is cured, 0 otherwise

Bernoulli function of P(Y_i|T_i,S_i)

W_i

Unmeasured patient factors causing variation in Y_i given T_i and S_i