Validation of the NASA Integrated Medical Model: a Space Flight Medical Risk Prediction Tool

Myers, Jerry G ; Garcia, Yamil Lionel ; Arellano, John D ; Boley, Lynn Addy ; Goodenow, Debra A ; Kerstman, Eric Lee ; Koslovsky, M ; Reyes, David Paul ; Saile, Lynn ; Taiym, Wafa F(Wyle Labs, Inc, Houston, TX, United States)

关键词: RISK;    SPACE TRANSPORTATION SYSTEM;    DATA ACQUISITION;    INTERNATIONAL SPACE STATION;    MONTE CARLO METHOD;    PROBABILITY THEORY;    SIGNS AND SYMPTOMS;    SPACE SHUTTLE MISSIONS;    AEROSPACE MEDICINE;    MODELS;   

RP-ID  :  GRC-E-DAA-TN60336

学科分类：医学（综合）

美国|英语

来源: NASA Technical Reports Server

【 摘 要 】

The Human Research Program funded the development of the Integrated Medical Model (IMM) to quantify the medical component of overall mission risk. The IMM uses Monte Carlo simulation methodology, incorporating space flight and ground medical data, to estimate the probability of mission medical outcomes and resource utilization. To determine the credibility of IMM output, the IMM project team completed two validation studies that compared IMM predicted output to observed medical events from a selection of Shuttle Transportation System (STS) and International Space Station (ISS) missions. The validation study results showed that the IMM underpredicted the occurrence of ~10% of the modeled medical conditions for the STS missions and overpredicted ~20% of the modeled medical conditions for the ISS missions. These findings imply that the strength of IMM predictions to inform decisions depends on simulated mission specifications including length. This discrepancy could result from medical recording differences between ISS and STS that possibly influence observed incidence rates, IMM combining all "mission type" data as constant occurrence rate or fixed proportion across both mission types, misspecification of symptoms to conditions, and gaps in the literature informing the model. Some of these issues will be alleviated by updating the IMM source data through incorporation of the observed validation data.