Resources

A bioethics review examines commercial Direct-to-Consumer (DTC) polygenic risk scores (PRS) for type 2 diabetes. In clinical genetics, odds ratios are defined using the disease-risk allele as the referent (OR>1.00OR > 1.00). However, an unaccredited commercial service inverted its reference coordinate axis, defining OROR relative to the protective allele and introducing a severe sign/directionality trap. For an individual carrying a risk genotype with a true validated odds ratio of ORtrue=1.80OR_{\text{true}} = 1.80, the software erroneously logged the log-odds effect size as βreported=−ln⁡(ORtrue)\beta_{\text{reported}} = -\ln(OR_{\text{true}}). (a) Calculate the numerical value of βreported\beta_{\text{reported}} to three decimal places, and (b) calculate the falsely inverted relative risk odds ratio ORreported=eβreportedOR_{\text{reported}} = e^{\beta_{\text{reported}}} presented to the patient, explaining the resulting clinical misclassification.

Answer and steps

Final answer

(a) βreported=−ln⁡(1.80)≈−0.588\beta_{\text{reported}} = -\ln(1.80) \approx -0.588; (b) ORreported=e−0.588≈0.556OR_{\text{reported}} = e^{-0.588} \approx 0.556; the patient is falsely informed that their genotype confers a 44.4%44.4\% reduction in odds of developing diabetes instead of an 80.0%80.0\% increase.

Steps

Part (a): Calculating the Erroneously Reported Log-Odds Effect Size (βreported\beta_{\text{reported}})

In quantitative genetic association studies, the effect size β\beta of an allele is given by the natural logarithm of its odds ratio (OROR):

β=ln⁡(OR)\beta = \ln(OR)

When the software flips the reference allele from the protective allele to the risk allele (or vice versa), it inverts the sign of the effect size:

βreported=−ln⁡(ORtrue)\beta_{\text{reported}} = -\ln(OR_{\text{true}})

Given that the true odds ratio is ORtrue=1.80OR_{\text{true}} = 1.80:

βreported=−ln⁡(1.80)≈−0.5877866…\beta_{\text{reported}} = -\ln(1.80) \approx -0.5877866\dots

Rounding to three decimal places:

βreported≈−0.588\beta_{\text{reported}} \approx -0.588


Part (b): Calculating ORreportedOR_{\text{reported}} and Evaluating Clinical Misclassification

  1. Calculate the reported odds ratio (ORreportedOR_{\text{reported}}):

    Using the exponential function to convert the reported log-odds back to an odds ratio:

    ORreported=eβreported=e−ln⁡(1.80)=11.80≈0.5555…OR_{\text{reported}} = e^{\beta_{\text{reported}}} = e^{-\ln(1.80)} = \frac{1}{1.80} \approx 0.5555\dots

    Rounding to three decimal places:

    ORreported≈0.556OR_{\text{reported}} \approx 0.556

  2. Evaluate the clinical and ethical implications:

    • True clinical reality: An odds ratio of ORtrue=1.80OR_{\text{true}} = 1.80 means the individual faces an 80.0%80.0\% increase in the odds of developing type 2 diabetes relative to average population risk: (1.80−1.00)×100%=+80.0%(1.80 - 1.00) \times 100\% = +80.0\%
    • Erroneously reported outcome: An odds ratio of ORreported≈0.556OR_{\text{reported}} \approx 0.556 falsely reports that the individual has a protective genotype, corresponding to a 44.4%44.4\% decrease in disease odds: (1.00−0.556)×100%=44.4% reduction(1.00 - 0.556) \times 100\% = 44.4\% \text{ reduction}
    • Impact: The patient receives dangerous false reassurance. Believing they carry substantial genetic protection against diabetes, they may forego preventative diet and exercise regimens or delay recommended clinical screenings. This highlights critical regulatory concerns regarding the validation, clinical oversight, and quality control of direct-to-consumer (DTC) genomic testing platforms.

Typical errors

OverviewACEpath Teacher's Diagnosis

01 A student might confuse the natural logarithm with the common base-10 logarithm when computing the log-odds effect size.

Sample Student Work

Reviewed by ACEpath Teacher
Step-by-Step Review
Walk Through It with Your Teacher
Partially correct
  1. Step 1

    βreported=−log⁡10(1.80)  ⟹  βreported=−0.255\beta_{\text{reported}} = -\log_{10}(1.80) \implies \beta_{\text{reported}} = -0.255

    Correct

    βreported=−ln⁡(1.80)≈−0.588\beta_{\text{reported}} = -\ln(1.80) \approx -0.588

    You used the common logarithm log⁡10\log_{10} rather than the natural logarithm ln⁡\ln. In genetics and logistic regression, log-odds effect sizes are defined using the natural logarithm with base ee.

  2. Step 2

    ORreported=e−0.255≈0.775OR_{\text{reported}} = e^{-0.255} \approx 0.775

    Correct

    ORreported=e−0.588≈0.556OR_{\text{reported}} = e^{-0.588} \approx 0.556

    Your formula eβreportede^{\beta_{\text{reported}}} was conceptually correct, but your result was affected by the erroneous β\beta value from Step 1.

  3. Step 3

    Since ORreportedOR_{\text{reported}} is less than 1.001.00, the patient is incorrectly informed that they have lower risk rather than higher risk.

WHAT YOU DID WELL
Compliments from Your Teacher

You correctly applied the exponential function eβe^\beta to convert log-odds back into an odds ratio.

You accurately identified that an odds ratio below 1.001.00 represents a falsely reported decrease in clinical disease risk.

THE ERROR
Where Your ACEpath Teacher Spotted the Mistake
Logarithm Base Confusion

no

WHY IT HAPPENED
ACEpath Teacher's Diagnosis

You substituted the base-10 logarithm (log⁡10\log_{10}) instead of the natural logarithm (ln⁡\ln or log⁡e\log_e) when evaluating the effect size β\beta. In epidemiological and genetic statistics, odds ratios relate to logit parameters via the natural exponential and logarithmic functions.

REASONING PATH
Your ACEpath teacher’s suggested approach
Natural Logarithm vs. Common Logarithm
  • In genetics and biostatistics, log-odds effect sizes are always defined as β=ln⁡(OR)=log⁡e(OR)\beta = \ln(OR) = \log_e(OR), which corresponds inversely to OR=eβOR = e^\beta.

  • Ensure that your calculator is set to evaluate ln⁡\ln (base ee) rather than log⁡\log (base 1010) whenever working with logistic regression or log-odds.

CORRECTED PATH
ACEpath Teacher's Demonstration
  1. Compute βreported=−ln⁡(1.80)≈−0.588\beta_{\text{reported}} = -\ln(1.80) \approx -0.588.

  2. Calculate the reported odds ratio: ORreported=e−0.588=11.80≈0.556OR_{\text{reported}} = e^{-0.588} = \frac{1}{1.80} \approx 0.556.

Final answer

(a) βreported≈−0.588\beta_{\text{reported}} \approx -0.588; (b) ORreported≈0.556OR_{\text{reported}} \approx 0.556; the patient is falsely told they have a 44.4%44.4\% reduction in odds instead of an 80.0%80.0\% increase.

NEXT STEPS
Final Takeaways from Your Teacher
  • Double-check that statistical conversions between odds ratios and effect sizes use base ee (ln\\ln) exclusively.

  • Practice converting between odds ratios and beta\\beta coefficients in both directions to solidify the natural exponential relationship.

OverviewACEpath Teacher's Diagnosis

02 A student might omit the negative sign in the given definition of the erroneously logged log-odds effect size.

Sample Student Work

Reviewed by ACEpath Teacher
Step-by-Step Review
Walk Through It with Your Teacher
Incorrect
  1. Step 1

    βreported=ln⁡(1.80)  ⟹  βreported=0.588\beta_{\text{reported}} = \ln(1.80) \implies \beta_{\text{reported}} = 0.588

    Correct

    βreported=−ln⁡(1.80)≈−0.588\beta_{\text{reported}} = -\ln(1.80) \approx -0.588

    You omitted the negative sign from the problem's given formula βreported=−ln⁡(ORtrue)\beta_{\text{reported}} = -\ln(OR_{\text{true}}), resulting in a positive log-odds value of +0.588+0.588 instead of −0.588-0.588.

  2. Step 2

    ORreported=e0.588=1.80OR_{\text{reported}} = e^{0.588} = 1.80

    Correct

    ORreported=e−0.588=11.80≈0.556OR_{\text{reported}} = e^{-0.588} = \frac{1}{1.80} \approx 0.556

    You correctly exponentiated your value from Step 1, but because βreported\beta_{\text{reported}} was missing its negative sign, this returned the un-inverted odds ratio of 1.801.80 rather than the erroneous reported ratio of 0.5560.556.

  3. Step 3

    The patient is correctly classified as having elevated risk because the inversion was not applied.

    Correct

    The patient is falsely told their odds ratio is 0.5560.556 (a 44.4%44.4\% decrease in disease odds instead of an 80.0%80.0\% increase), leading to dangerous false reassurance.

    Because the inversion sign was omitted in your calculation, you concluded the patient was correctly classified. Evaluating the software's actual error shows that reporting OR=0.556OR = 0.556 falsely classifies a high-risk patient as having genetic protection.

WHAT YOU DID WELL
Compliments from Your Teacher

You correctly evaluated ln⁡(1.80)≈0.588\ln(1.80) \approx 0.588 to three decimal places.

You understood that recovering an odds ratio from log-odds requires the exponential function eβe^{\beta}.

THE ERROR
Where Your ACEpath Teacher Spotted the Mistake
Omission of negative sign in coordinate inversion

Dropping the leading negative sign in the software error definition βreported=−ln⁡(ORtrue)\beta_{\text{reported}} = -\ln(OR_{\text{true}}).

WHY IT HAPPENED
ACEpath Teacher's Diagnosis

Conflating the true biological effect size βtrue=ln⁡(ORtrue)\beta_{\text{true}} = \ln(OR_{\text{true}}) with the software's flawed inverted parameter βreported=−ln⁡(ORtrue)\beta_{\text{reported}} = -\ln(OR_{\text{true}}), which bypassed the prompt's scenario of an inverted reference allele.

REASONING PATH
Your ACEpath teacher’s suggested approach
Check Allele Reference Orientation
  • Flipping the reference allele from risk to protective inverts the odds ratio (ORinverted=1OROR_{\text{inverted}} = \frac{1}{OR}) and negates log-odds (βinverted=−β\beta_{\text{inverted}} = -\beta).

  • Before calculating, verify whether an effect size represents risk (eta > 0, OR > 1) or protection (eta < 0, OR < 1).

CORRECTED PATH
ACEpath Teacher's Demonstration
  1. Apply the negative sign given for the inverted coordinate axis: βreported=−ln⁡(1.80)≈−0.588\beta_{\text{reported}} = -\ln(1.80) \approx -0.588.

  2. Compute the reported odds ratio: ORreported=e−0.588=11.80≈0.556OR_{\text{reported}} = e^{-0.588} = \frac{1}{1.80} \approx 0.556.

  3. Interpret the clinical impact: An OROR of 0.5560.556 indicates a (1−0.556)×100%=44.4%(1 - 0.556) \times 100\% = 44.4\% reduction in diabetes odds, falsely reassuring a patient whose true risk is elevated by 80.0%80.0\%.

Final answer

(a) βreported≈−0.588\beta_{\text{reported}} \approx -0.588; (b) ORreported≈0.556OR_{\text{reported}} \approx 0.556. The patient is falsely informed of a 44.4%44.4\% risk reduction instead of an 80.0%80.0\% risk increase, causing dangerous false reassurance.

NEXT STEPS
Final Takeaways from Your Teacher
  • Reread problem definitions carefully to ensure prescribed negative signs or directional inversions are carried into your initial equation.

  • Practice translating odds ratios across 1.001.00: identify values >1.00> 1.00 as increased risk and values <1.00< 1.00 as decreased risk/protective.

OverviewACEpath Teacher's Diagnosis

03 A student might correctly calculate the inverted odds ratio but misinterpret an odds ratio below 1.00 as indicating no effect rather than a protective effect.

Sample Student Work

Reviewed by ACEpath Teacher
Step-by-Step Review
Walk Through It with Your Teacher
Partially correct
  1. Step 1

    βreported=−ln⁡(1.80)  ⟹  βreported=−0.588\beta_{\text{reported}} = -\ln(1.80) \implies \beta_{\text{reported}} = -0.588

  2. Step 2

    ORreported=e∧(−0.588)≈0.556OR_{\text{reported}} = e^{\wedge}(-0.588) \approx 0.556

  3. Step 3

    Because the value is below 1.001.00, the patient is misclassified as having zero risk or no association with the disease instead of a protective effect.

    Correct

    An odds ratio of ORreported≈0.556OR_{\text{reported}} \approx 0.556 represents a 44.4%44.4\% reduction in odds (a protective effect). Since the patient actually carries a risk genotype with ORtrue=1.80OR_{\text{true}} = 1.80 (80.0%80.0\% increased odds), the patient is falsely reassured that they are genetically protected rather than at elevated risk.

    An odds ratio of 1.001.00 corresponds to the null value (no association), while values strictly between 00 and 1.001.00 indicate reduced odds or a protective effect. More critically, the true clinical baseline is elevated risk (OR=1.80OR = 1.80), so the software erroneously reports protection instead of an 80%80\% risk increase.

WHAT YOU DID WELL
Compliments from Your Teacher

You correctly calculated the negative log-odds effect size to three decimal places.

You correctly exponentiated the reported beta value to find the inverted odds ratio.

THE ERROR
Where Your ACEpath Teacher Spotted the Mistake
odds_ratio_interpretation_and_null_value_confusion

no

WHY IT HAPPENED
ACEpath Teacher's Diagnosis

You confused the null reference point of an odds ratio (OR=1.00OR = 1.00) with values indicating a protective effect (OR<1.00OR < 1.00), and you overlooked that the patient's true genotype confers increased risk (OR=1.80OR = 1.80) rather than being compared against an intended protective baseline.

REASONING PATH
Your ACEpath teacher’s suggested approach
Odds Ratio Scale Interpretation
  • On a multiplicative scale, OR=1.00OR = 1.00 means no association (equal odds), OR>1.00OR > 1.00 indicates increased risk, and 0<OR<1.000 < OR < 1.00 indicates reduced risk or protection.

  • Always treat 1.001.00 as the null baseline and convert OR<1.00OR < 1.00 to percent risk reduction using (1.00−OR)×100%(1.00 - OR) \times 100\%.

Clinical Directionality Comparison
  • Clinical misclassification requires contrasting the reported classification against the true biological status (ORtrueOR_{\text{true}}).

  • Explicitly state what the true genotype causes (ORtrue=1.80  ⟹  +80%OR_{\text{true}} = 1.80 \implies +80\% odds) versus what the erroneous report tells the patient (ORreported=0.556  ⟹  −44.4%OR_{\text{reported}} = 0.556 \implies -44.4\% odds), noting the resulting false reassurance.

CORRECTED PATH
ACEpath Teacher's Demonstration
  1. Recognize that OR=1.00OR = 1.00 denotes no association, whereas OR=0.556OR = 0.556 means the patient's odds are reduced to 55.6%55.6\% of baseline (a (1.00−0.556)×100%=44.4%(1.00 - 0.556) \times 100\% = 44.4\% reduction).

  2. Compare this reported protective effect to the patient's true risk status of ORtrue=1.80OR_{\text{true}} = 1.80, which represents an (1.80−1.00)×100%=80.0%(1.80 - 1.00) \times 100\% = 80.0\% increase in disease odds.

  3. State the clinical misclassification: the patient is falsely given reassurance of genetic protection against type 2 diabetes when they are actually at significantly elevated risk, potentially leading them to neglect necessary lifestyle interventions or screenings.

Final answer

(a) βreported≈−0.588\beta_{\text{reported}} \approx -0.588; (b) ORreported≈0.556OR_{\text{reported}} \approx 0.556; the patient is falsely told their genotype provides a 44.4%44.4\% reduction in diabetes odds (protective) instead of conferring an 80.0%80.0\% increase in odds (elevated risk), creating dangerous false reassurance.

NEXT STEPS
Final Takeaways from Your Teacher
  • Review the difference between the null value for log-odds (β=0\beta = 0) and the null value for odds ratios (OR=1.00OR = 1.00).

  • Practice calculating the percentage change in risk from odds ratios both above and below 1.001.00.

Lock it in

Similar question

01

A genetic association study reports an effect size of β=0.405\beta = 0.405 for an allele. Calculate the odds ratio (OROR) associated with this allele to two decimal places.

Answer

OR=e0.405≈1.50OR = e^{0.405} \approx 1.50. The allele confers a 50%50\% increase in disease odds.

02

A direct-to-consumer software bug inverts the reference allele for a variant with true odds ratio ORtrue=2.50OR_{\text{true}} = 2.50, reporting βreported=−ln⁡(2.50)\beta_{\text{reported}} = -\ln(2.50). Calculate the reported odds ratio ORreportedOR_{\text{reported}} and state the percentage change in disease odds falsely communicated to the patient.

Answer

βreported=−ln⁡(2.50)≈−0.916\beta_{\text{reported}} = -\ln(2.50) \approx -0.916. Thus ORreported=e−0.916=12.50=0.400OR_{\text{reported}} = e^{-0.916} = \frac{1}{2.50} = 0.400. The patient is falsely informed of a 60.0%60.0\% decrease in odds (1.00−0.400=0.6001.00 - 0.400 = 0.600) instead of a 150%150\% increase.

03

A polygenic risk score aggregates two independent single-nucleotide polymorphisms with true odds ratios OR1=1.25OR_1 = 1.25 and OR2=1.60OR_2 = 1.60. If a laboratory accidentally computes log-odds using base-10 logarithms (βi=log⁡10(ORi)\beta_i = \log_{10}(OR_i)) but exponentiates the sum using base ee (ORcombined=eβ1+β2OR_{\text{combined}} = e^{\beta_1 + \beta_2}), calculate the erroneous combined odds ratio and compare it to the true combined odds ratio.

Answer

True combined odds ratio: ORtrue=OR1×OR2=1.25×1.60=2.00OR_{\text{true}} = OR_1 \times OR_2 = 1.25 \times 1.60 = 2.00 (since βtrue=ln⁡(1.25)+ln⁡(1.60)=ln⁡(2.00)≈0.693\beta_{\text{true}} = \ln(1.25) + \ln(1.60) = \ln(2.00) \approx 0.693). Erroneous sum: βerr=log⁡10(1.25)+log⁡10(1.60)=log⁡10(2.00)≈0.30103\beta_{\text{err}} = \log_{10}(1.25) + \log_{10}(1.60) = \log_{10}(2.00) \approx 0.30103. Erroneous reported odds ratio: ORreported=e0.30103≈1.35OR_{\text{reported}} = e^{0.30103} \approx 1.35. The patient's genetic risk is substantially underestimated (1.351.35 vs. 2.002.00).

Loading workspace…