CAT 3 Harm, Causation Maribeth Chitkara, MD Rachel Boykan, MD

Harm or Causation

Answerable Clinical Question You are the intern rounding in the newborn nursery. You are discussing your newest admission on the phone with your attending. The baby is a healthy term girl; however, her mother tested positive for chlamydia upon this admission. You are wondering if you should start the baby on oral erythromycin. Your attending thinks it’s a good idea, but she wants you to discuss the risk of developing pyloric stenosis with the baby’s mother. You are unaware of any association of erythromycin and pyloric stenosis, so you hit the computer to do some research. You are the intern rounding in the newborn nursery. You are discussing your newest admission on the phone with your attending. The baby is a healthy term girl; however, her mother tested positive for chlamydia upon this admission. You are wondering if you should start the baby on oral erythromycin. Your attending thinks it’s a good idea, but she wants you to discuss the risk of developing pyloric stenosis with the baby’s mother. You are unaware of any association of erythromycin and pyloric stenosis, so you hit the computer to do some research.

Study Designs for Causation DesignStarting point AssessmentStrengthsWeakness Randomized Controlled Trial ExposureAdverse event Internal validityEthics, feasibility CohortExposureAdverse event Feasible when randomization not possible Internal validity Case ControlAdverse eventExposureOvercomes time issues, may only require small sample size Internal validity Case ReportAdverse eventExposureStarting pointGeneralizability

Harm/Causation – Validity Comparison groups? Comparison groups? Clearly defined and similar groups? Clearly defined and similar groups? Assessment of outcomes objective or blinded? Assessment of outcomes objective or blinded? Follow up long enough? Follow up long enough? Fulfills diagnostic tests for causation? Fulfills diagnostic tests for causation?

Association: - Harm or Etiology Validity Were there clearly defined groups of patients, similar in all important ways other than exposure to the treatment / cause? Were there clearly defined groups of patients, similar in all important ways other than exposure to the treatment / cause? Was the assessment of outcomes either objective or blinded to the exposure? Was the assessment of outcomes either objective or blinded to the exposure? Was the follow-up of the study patients sufficiently long (for the outcome to occur) and complete? Was the follow-up of the study patients sufficiently long (for the outcome to occur) and complete? Do the results of the harm study fulfill some of the diagnostic tests for causation? Do the results of the harm study fulfill some of the diagnostic tests for causation?

Association: Harm or Etiology Were the exposures and outcomes measured the same way in both groups? Were the exposures and outcomes measured the same way in both groups? How was the exposure ascertained? How was the exposure ascertained? Recall Bias or Interviewer Bias Recall Bias or Interviewer Bias Exposure Opportunity Exposure Opportunity How was the outcome ascertained? How was the outcome ascertained? Surveillance Bias Surveillance Bias Look for strategies that minimize bias, such as blinding the subjects and interviewers to the hypothesis of the study Look for strategies that minimize bias, such as blinding the subjects and interviewers to the hypothesis of the study

Association: Harm or Etiology Diagnostic Tests for Causation Is it clear that the exposure preceded the onset of the outcome? Is it clear that the exposure preceded the onset of the outcome? Is there a dose-response gradient? Is there a dose-response gradient? Is there any positive evidence from a de-challenge / re-challenge study? Is there any positive evidence from a de-challenge / re-challenge study? Is the association consistent from study to study? Is the association consistent from study to study? Does the association make biological sense? Does the association make biological sense?

Causation / Harm – Results Absolute Risk Increase (ARI) = c / c+d – a / a+b

Association: Harm or Etiology What is the Magnitude of the Risk? ARI = Absolute Risk Increase ARI = Absolute Risk Increase ARI = a / a+b -c / c+d ARI = a / a+b -c / c+d NNH (Number needed to HARM) = 1 / ARI NNH (Number needed to HARM) = 1 / ARI

Association: Harm or Etiology Relative Risk (RR) Disease Prevalence in an exposed vs. non-exposed population Disease Prevalence in an exposed vs. non-exposed population The risk (or incidence) of the adverse effect in the exposed group divided by the risk of the adverse effect in the unexposed group The risk (or incidence) of the adverse effect in the exposed group divided by the risk of the adverse effect in the unexposed group Actual prevalence data needed for 2x2 chart Actual prevalence data needed for 2x2 chart Describes the disease prevalence in specific exposed vs. non- exposed populations Describes the disease prevalence in specific exposed vs. non- exposed populations RR > 1 represent an increase in risk associated with the exposure RR > 1 represent an increase in risk associated with the exposure

Association: Harm or Etiology Relative Risk (RR) RR depends on having actual data on exposed and unexposed patients in a designated population RR depends on having actual data on exposed and unexposed patients in a designated population It is not applicable in Case-Control studies, where the number of cases and controls is chosen by the investigator It is not applicable in Case-Control studies, where the number of cases and controls is chosen by the investigator For C-C studies, the Odds Ratio is used For C-C studies, the Odds Ratio is used

Association: Harm or Etiology Odds Ratio (OR) Ratio of Odds Ratio of Odds The odds of a case patient being exposed divided by the odds of a control patient being exposed The odds of a case patient being exposed divided by the odds of a control patient being exposed Proportion exposed in a diseased vs. non- diseased patient sample Proportion exposed in a diseased vs. non- diseased patient sample When the outcome of interest is rare in the population from which the sample was drawn (often the reason for using a case-control study), the OR closely approximates the RR When the outcome of interest is rare in the population from which the sample was drawn (often the reason for using a case-control study), the OR closely approximates the RR

Association: Harm or Etiology Odds Ratio (OR) Describes the relative harm of an exposure independent of disease prevalence Describes the relative harm of an exposure independent of disease prevalence OR = 1 No effect OR > 1+ Harm If the disease prevalence is small, then RR~OR If the disease prevalence is small, then RR~OR

Association: Harm or Etiology  Examine the confidence interval around the estimate of risk  Larger confidence intervals make the result less significant  Confidence intervals that cross 1 (null) make the result less significant

Association: Harm or Etiology What is the Magnitude of the Risk? The RR and the OR do not tell us how frequently a problem occurs, only that the effect occurs more or less often in the exposed group compared with the unexposed group

Harm or Etiology Are the results applicable to my patient? Is my patient so different from those in the study that the results don’t apply? Is my patient so different from those in the study that the results don’t apply? What is my patient’s risk of an adverse event? What is my patient’s risk of an adverse event? What is my patient’s potential benefit from the therapy? What is my patient’s potential benefit from the therapy? What are my patient’s preferences, concerns, and expectations from this treatment? What are my patient’s preferences, concerns, and expectations from this treatment? What alternative treatments are available? What alternative treatments are available?