Unmet Need for Massive Transfusion in Indian Trauma Patients

Overview

This study aims to estimate the difference between predicted and provided massive transfusion in adult trauma patients admitted to university hospitals in urban India.

Background Trauma fatalities account for 9 % of the world's deaths, superseding the number of deaths caused by HIV/AIDS, tuberculosis and malaria combined, disproportionately affecting low- and middle-income countries (LMIC). Hemorrhage is one of the leading causes of preventable death among trauma patients. Up to 5 % of trauma patients require massive transfusions (MT), defined as administration of ten or more units of packed red blood cells (PRBC) within 24 hours of injury. India is an LMIC lacking structured trauma care systems and where blood products remain a scarce commodity. There is a sizeable gap in supply and demand as India lacks about four million units of blood per year according to World Health Organization estimates. The in-hospital mortality in an urban Indian trauma setting has been found to be twice that of high-income equivalents and inadequate resuscitation and bleeding control of hemorrhaging patients has been found to be the main cause of preventable deaths. Considering the general lack of blood components in India it is reasonable to assume that there is an unmet need of blood transfusions in trauma patients as well. The extent of the potentially unmet need of MT constitutes a major gap of knowledge and also an important basis for resource allocation and development of MT protocols suited for local needs, hence our research question: what is the difference between predicted need for MT and actual transfusion in an urban Indian trauma setting? Aim To estimate the difference between predicted need for MT and provided MT in adult trauma patients in an urban Indian setting. The hypothesis is that there will be a substantial discrepancy between predicted need and provided MT, indicating an unmet need of blood transfusions. Study design A retrospective analysis of the Towards Improved Trauma Outcomes in India (TITCO) cohort will be conducted. Setting The TITCO dataset includes 16,047 patients enrolled between July 2013 and December 2015 from four public university hospitals, namely the King Edward Memorial Hospital and Lokmanya Tilak Municipal General Hospital in Mumbai, Jai Prakash Narayan Apex Trauma Center in New Delhi and the Seth Sukhlal Karnani Memorial Hospital in Kolkata. These hospitals are characterized by nominal patient fees and all operate trauma units. An externally employed and trained data collector working eight-hour shifts in each facility gathered patient data in a prospective observational fashion, rotating between day, evening and night shifts. Data from patients admitted outside of these hours was retrieved retrospectively within days of admission. The patients were followed until discharge or death, whichever occurred first. Source and method of participant selection Participants were selected consecutively during the study period according to the aforementioned eligibility criteria, either via direct observation during the data collector's shift or through retrospective data retrieval from patient records. Covariates To estimate the proportion of patients predicted to need MT the Assessment of Blood Consumption (ABC) score will be used. The rationale for using the ABC score in this context is that it is based on routinely recorded vital signs and examination results that are all available in TITCO, unlike other scores that also require laboratory results. Further, the intent is not to validate the score, but to use it as an informed basis for identifying patients with substantial bleeding. ABC is based on four dichotomous non-laboratory parameters that are readily available in the early assessment phase. Penetrating mechanism (0=no, 1=yes) Arrival Systolic Blood Pressure \<= 90 mm Hg (0=no, 1=yes) Arrival Heart Rate \>= 120 bpm (0=no, 1=yes) Positive focused assessment sonography for trauma (FAST) (0=no, 1=yes) A score of two or more has been set to indicate need for MT. Age, sex, and mechanism of injury in addition to the variables included in the ABC score will be presented to characterize the study sample. Data sources/measurement Where FAST recording in the study sample is not reported as "positive" relevant free text terms will be used to find indications of free fluid in abdomen, pelvis or pericardium. Where FAST recording is missing relevant ICD-10 codes will be used for screening the free text variables reporting computer tomography (CT) and intraoperative findings. The proportion of patients predicted to need MT (a) will be estimated by dividing the number of patients with an ABC score of two or more by the total number of patients in the study sample. The proportion of actual MT received (b) will then be calculated by dividing the number of patients who received ten or more units of PRBC with the total number of patients with complete clinical data. Bias The data collectors were externally trained and funded by project management. The data was obtained through direct observation and retrieval from patient records without interference in clinical work. All data collectors had at least a health science master's degree and were continuously trained and supervised by project management. The collected data is multi-institutional, likely reducing the effects of highly operator and setting dependent variables. Injury severity bias due to missing data from severely injured patients is a recognized problem in trauma registries and to counteract this it is imperative to account for as many patients as possible using alternatives for the FAST positive variable that would indicate ongoing intra-abdominal, pericardial or pelvic hemorrhage. However, the approach of using surrogates for FAST may also yield misleading results due to overestimation. To estimate the magnitude of this source of bias a sensitivity analysis will be conducted. In addition, the ABC score was originally developed and validated in a high-income setting with a likelihood of differing demography and patterns of injury mechanism, possibly yielding results not representable of true MT needs in our study population. However, in this case the ABC score will rather be used as a tool for the estimation of blood needs rather than as a prediction tool or aid in clinical decision-making. Study size Based on previous research it is assumed that 10% of patients will be predicted to need MT. In contrast, based on clinical experience from the study setting, it is assumed that a maximum of 5% received it. Assuming that 10 % (\~ 1600 patients) will be predicted to need an MT and 5 % (\~ 800 patients) received it then the sample size based on a two-sided paired comparison of proportions, i.e. two-sided McNemar's test, needed to detect this difference with 80% power is 369 patients, setting alpha to 5% and beta to 20%. In other words, because all complete observations in the cohort will be analyzed, it can safely be said that this study is powered to detect this difference at the 80% level. Quantitative variables When characterizing the study sample, all quantitative variables will be presented as continuous. In the calculation of the ABC score the quantitative variables arrival heart rate and arrival systolic blood pressure will be dichotomized as detailed in the covariate section above. Statistical methods All statistical analyses will be performed in R, a free open-source software for statistical computing and graphics. Qualitative sample characteristics will be presented using number and percentage whereas quantitative sample characteristics will be presented as median, values at the 25th and 75th percentile, as well as minimum and maximum values. The main analysis is a complete case analysis where observations with missing covariate values will be excluded. Depending on the amount of missingness, multiple imputation using chained equations might be used to handle missing data. The main analysis will thereafter be conducted in two steps. First, the proportion of patients with an ABC score ≥ 2, i.e. patients with predicted need of MT will be calculated. Thereafter the proportion of patients who actually received MT (b) will be calculated based on transfusion data in the TITCO registry. The outcome is defined as the difference between the proportion of adult patients predicted to receive MT and the proportion of patients who did receive MT. Sensitivity analyses Based on clinical experience in this particular setting a sensitivity analysis will be conducted with a cutoff value of 4 units of PRBCs in 24 hours for the definition of MT. A sensitivity analysis will also be conducted by restricting the analysis to FAST positive patients where coding using CT or intraoperative findings was not used.

Conditions

Interventions

Eligibility

Outcomes