Statistics in Medicine
Evaluating medications, treatments, and procedures.
Today we've come to expect the certainties in life. We know so much, and can do so much, that the existence of uncertainties baffles us. This is especially true in medicine, where lives are quite literally on the line.
However, there are a number of challenges in the application of statistics in medicine, some of which also apply to other fields, others that are specific to medicine. There is a serious lack of data due to the lack of instances, funding to reproduce results, and inconsistencies between different subjects. This typically results in high variance and dependence on further research/replications by peers. However, statistics is not only used to get definite answers like, "how much of x should be used for person y that has disease z?" Statistics is also used to measure the directionality of the effect, suggest types of data to be collected, improve the process in which the research is conducted, and serve as an interpretable baseline for which future works can be built on.
Below are examples of such works.
Cancer Treatments: Bevacizumab and PD901
Bevacizumab, otherwise known as Avastin, is a medication used to treat a variety of types of cancers. In the study conducted by the client (Prof. Mark Conaway), an experiment using 4 doses of bevacizumab {0, 0.05, 0.10, and 0.25 mm} and 4 doses of PD901 {0, 0.005, 0.01, and 0.025 mm} were run with 3 biological replicates. The main objective of this statistical study is to determine whether bevacizumab and PD901 act additively in inhibiting cell growth.
Analysis of the client's dataset uncovered noteworthy results. Statistical modeling supports that bevacizumab and PD901 do act additively in inhibiting cell growth. However, the additive effects are not constant. The synergy of the two drugs grows strong or weaker based on the different combinations of dosage.
Strength Training as Treatment for Cerebral Palsy
Cerebral palsy is a non-progressive neurological disorder that primarily affects body movement and muscle coordination. Approximately two out of every thousand newborn children develop CP and to this date, there is no cure. This report discusses the relationship between the lower-body strength of the child with CP and their walking speed. As the risk involved with the strength-training treatment is low if not non-existent with proper care, even the slightest positive significant correlation should suggest the value of strength training.
Based on the study, there appears to be a strong relationship between strength and walking speed. This holds especially for those with Gross Motor Function Classification System (henceforth "GMFCS") Level 1, whether the CP is diplegic or hemiplegic. However, the strength of the relationship decays drastically for increasing GMFCS Levels. For those with GMFCS Levels of 3 or higher, possible benefits of the strength training appears to be negligible.
Evaluation of E-ICUs
Due to the rising concerns of mortality rates of intensive care unit (ICU) patients in local/regional levels and the ineffectiveness of the current "Resource-Centered Care", new approaches, such as E-ICU, have been proposed. E-ICU is essentially an off-site communication center for individuals in the regions without sufficient resources to get medical assistance as well as receive evaluation for appropriate transfers. The current study aims to review the cost, efficiency, and quality of the measures provided by E-ICU to evaluate E-ICU as a possible alternative to the current system. The data available suggests that E-ICU promises better overall patient care and reduces mortality rare for the patients -enough to flag E-ICU as a candidate for a more rigorous study. However, it must be noted that there is a clear lack of data which greatly weakens any statistical methods applied. The convenience sampling based on the region due to the nature of how data collection methods were implemented further weakens the data and the consequent analysis.