ADS Capstone Chronicles Revised
25
individual differences, especially weight, sex, and age in people taking certain medicationsarerelatedto death and serious outcomes. For males, age increases with seriousness and weight decreases with seriousness, showing that in general, older age and lower weights are associated with more serious outcomes. However, females show a different pattern. Age increasesand weight decreases from nonserious to serious , but for death , younger, overweight females are at risk (Figure 8a,b). Females with death outcome have similar younger ages to males experiencing nonserious events ( p = .12), maleswith death outcome have similar lower weights to females with nonserious outcome ( p = .79), and males with nonserious outcome have similar higher weight to females with death outcome ( p = .38). Asforthepredictordrugfeatures,thehigher splitsintherandomforestmodelweredrugs related to nonserious outcomes, which makes sense since 72% of nonserious outcomes are related toasingledrugcalled Dupixent which treats a variety of inflammatory conditions in the sinuses, lungs,andskin. Serious and death outcomes are related to a wider range of drugs, with the top reported representing only27%and 24% of the sample, respectively. Thus, the childnodepatternsinthetreesbecomemore complexwiththevarietyofdrugsrelatedto serious and death outcomes. For serious outcome, most drugs are forms of chemotherapy. For death outcome, the drugsareindicatedforinflammation,cancer treatment, and pain relief. The risk factors in the model suggest that differences in pharmacokinetics are related to adverse outcomes. Pharmacokinetics is the movement and metabolism of drug compounds in a living organism. Obesity affects the pharmacokinetics of drug compoundsbyslowingdrugabsorptionrates
and metabolism based on higher fat mass, and lower water and lean body mass, thus, on average, obese patients should be given smaller doses of drugscomparedtohealthy individuals with more lean body mass (Gouju & Legeay, 2023). This puts obese individuals at risk of overdose and ADRs. Females are at greater risk of death related to opioid, antipsychotic, benzodiazepine, andantidepressantmedicationscomparedto males, possibly due to drug interactions fromlackofcommunicationbetweenfemale healthcare providers, or easier access to mental-health drugs compared to males (Iwanicki et al, 2015; Wightman et al., 2021). Additionally, there may be a gender gap in clinical trial research for certain compounds whereby recommendeddosages are based on primarily male data. Age affects pharmacokinetics because older individuals are exposed to more drugs overall,andcertainepigeneticsignalingand enzyme production affects the ability to metabolize drug compounds as people age (Duecker & Brockmoeller, 2019). The patterns between these risk factors can be nonlinear, which is reflected in the high performance of the decision tree based models and lower performanceinthelinear models, and thelackofcorrelationbetween numerical features (Figure 9). 6.2 Limitations This project has limitations resulting from thenatureofFAERSdataandtheFDAdata infrastructure, but which can be addressed with future analyses. The interpretation of the results cannot be extrapolated to the entire population of people using the drugs reported in FAERS. The conclusions and insights can only be applied to people who take the drugs reported in FAERS and who also had healthcare providersmakeareportforthem in FAERS. Thus, the results only apply to
175
Made with FlippingBook - Online Brochure Maker