Page 72 - Journal of Special Operations Medicine - Winter 2014

intention to sustain their medical skills. The third fac- November 8, 2011, no monetary compensation for par-
tor, perceived behavior control, assessed the respondent’s ticipation was allowed.
control beliefs and the influence of that control. Con-
trol beliefs assessed each individual’s perceived level of Ethics Statement
control on how they conduct their sustainment training, The research protocol for this survey was approved by
and influence of control beliefs assessed how that level of the USUHS Institutional Review Board prior to cognitive
control influenced their ability to conduct the behavior. testing and survey deployment. The survey maintained
complete respondent anonymity and informed consent
These constructs represented the foundation for assess- was obtained when the respondent began the first page
ing medics’ attitudes, perceived barriers, and self- efficacy, of the survey (beyond the consent page). The consent
and were operationalized by developing questions relat- page highlighted the purpose of the study, the risks and
ing to each component of the constructs. Answer choices benefits of participation, alternatives to participation,
for each question used Likert-type scales: 5-point scales the right to withdraw, and where to get more informa-
were used for unipolar responses, and a 7-point scale tion. No personally identifiable or health information
was used for bipolar answers. Three open-ended ques- was collected and respondents were allowed to skip
tions were included to allow respondents to provide un- questions if desired.
structured responses. The survey was constructed using
LimeSurvey, an online survey program hosted on the se- Statistical Analyses
cure servers at the Uniformed Services University of the Prior to analysis, data were screened for accuracy, miss-
Health Sciences (USUHS). 18 ing values, and normality. Descriptive statistics were
calculated using frequency for categorical variables.
The final survey instrument included 47 items, several of Subgroup comparisons using demographic information
which included subcategories of medical care correspond- were made based on participant characteristics (e.g.,
ing to SFMS critical training requirements and training Group assignment, deployment history, and time since
manual. Three general, open-ended questions provided initial training) to identify underlying relationships be-
7
respondents the opportunity to provide feedback in an tween measured variables. When appropriate, analysis
unstructured format. Demographic information was col- of variance for subgroup comparisons or Student t test
lected on respondent age, unit, length of military service, to compare means was performed. Presurvey power
length of service in special operations, time since complet- analysis concluded that with a 40% response rate, 95%
ing sustainment training, and deployment experience. confidence intervals (CIs) for descriptive statistics would
generate a width of 20.7% of one standard deviation
Initial Survey Validation (SD). For comparing means, a response rate of 40%
Following initial development, the survey instrument would still provide 80% power to find a difference of
was evaluated and revised through subject matter ex- 30% of one SD (standardized effect of 0.3) based on a t
pert review and cognitive testing with former SFMS. test for independent samples with two groups of similar
Through each stage in the process, survey questions size and a 5%, two-sided significance level. All analyses
were revised to improve clarity and understanding. were completed using SPSS 22.0 with statistical signifi-
cance level of p ≤ .05 (IBM Corporation, Armonk, NY,
Survey Implementation USA; http://www-01.ibm.com/software/analytics/spss/).
This survey targeted all 869 SFMS (both active duty
and National Guard) serving in the US Special Opera-
tions Command (USASOC). Former SFMS or SFMS Results
serving outside of USASOC were excluded. To recruit The survey had a total of 334 respondents, which re-
participants, a study invitation e-mail with a link to sulted in a 38.4% overall response rate. Approximately
the online survey was distributed through the USASOC 89% of all respondents completed the entire survey. The
medical support channels to all potential respondents. average respondent was 32 years old (SD: 4.9) and has
The invitation contained a description of the study and served in the Army for 10.5 years (SD: 4.5) and in Spe-
a statement highlighting the voluntary nature of partici- cial Forces for 5.4 years (SD: 3.4). Nearly 93% of the
pation. Reminder e-mails with a link to the survey were respondents had deployed at least once, with the average
distributed to all potential respondents 2 and 4 weeks having 4.2 deployments (SD: 3.5). The average respon-
later. The survey was closed at the end of 6 weeks. Data dent has been deployed in a combat zone for a total of 13
collection was performed electronically and stored on months (SD: 9.4), deployed in a noncombat zone for 6.5
secure USUHS servers behind password protection. In months (SD: 9.3), and had returned from his last deploy-
accordance with Department of Defense Instruction ment 9.1 months (SD: 11.6) prior to the survey. There
3216.02, Protection of Human Subjects and Adher- was no difference in these values between the different
ence to Ethical Standards in DoD-Supported Research, Special Forces Groups (SFG) except for 19th Group,

62 Journal of Special Operations Medicine Volume 14, Edition 4/Winter 2014