Public health data surveillance uses epidemiological inference "to provide information to advance prevention and treatment strategies and inform public health policies" (Hoge et al., 2008). The goal is to inform policy/decision makers of necessary interventions for improving population health. Over the past decade clinical practice has moved toward evidence-based work—now public health is moving toward evidence-based practice (EBP).
The EBP movement has advanced largely through technological improvements in electronic health data collection, management, analysis, and reporting. Figure 1 [PDF] demonstrates the overlapping public health and risk mitigation processes (Jones et al., 2000; Petruccelli & Knapik, 2006, used with permission).
Clinicians using EBP also monitor disease outbreak, evaluate intervention, and review injury data. Information about rates of chronic disease or disability in different populations helps in estimating the numbers and types of health services required for affected cohorts.
Public Health Surveillance
In the late 1990s the Institute of Medicine (IOM) published a series of reports outlining methods for evaluating the performance of public health programs. Figure 1 [PDF] includes the IOM's overall principles about how the process should work. For example, public health programs work best when carried out by a multidisciplinary partnership that includes subject matter experts (SMEs) in the areas of clinical practice, epidemiology, bio-statistics, and clinical and statistical data management.
At the U.S. Army Center for Health Promotion and Preventive Medicine (CHPPM), multidisciplinary SME teams apply the IOM's principles of public health program evaluation and continuous improvement in an evidence-based military public health practice of disease and injury prevention and risk mitigation. Over the past decade these scientists have used advanced technologies to improve methods that monitor medical/health data for disease and injury outbreaks requiring intervention in military populations.
Electronic health data programs, including the composite health care system and the ambulatory data module, speed up data acquisition, management, analysis, and reporting as shown in Figure 1 [PDF]. Two essential elements of information within these systems are the ICD-9CM diagnostic codes and the CPT procedure codes.
The clinician's role is to provide clinical outcome metrics associated with a particular public health program subject matter area. The data for different types of injuries (acoustic injury, eye injury, etc.) are viewed primarily with respect to the standard ICD-9CM coding practices of different injury specialty providers and coders inputting the codes into military electronic health systems.
The epidemiologist's role applies epidemiological inference, including rates of disease and injury metrics among different cohorts, to data. The biostatistician helps identify the most appropriate statistical designs for analysis. This clinical and statistical data management aspect includes SMEs working together on the additional data sets with which the medical data can be integrated for most accurate organization for multivariate analyses.
Other military data sets accessed for multi-source data fusion include the categories of personnel, health survey, and medical evacuation. As analyses are performed, the SMEs work together to offer the most likely interpretation of differences in prevalence of disease or injury among the different cohorts. They are also required to present intervention strategies that are reasonable and actionable.
Acoustic Injury Surveillance
The audiology, epidemiology, and biostatistics SMEs at the CHPPM have worked together to develop an acoustic injury (AcI) surveillance process that best fits in with the surveillance processes of other clinical specialties. These SMEs also have started specializing in surveillance methods development for postdeployment acoustic injury.
Since 2004 we have taken steps to improve and refine our postdeployment surveillance processes related to acoustic injury. Our intention is to implement this surveillance process with those of other injury specialties and as a separate postdeployment acoustic injury process. In 2006 several injury specialty programs at the center, including acoustic injury, were invited to participate in the Defense Safety Oversight Council's Military Injury Epidemiology and Prevention Working Group.
The council tasked the working group to evaluate medical data to find injury categories in which evidence-based interventions had the highest likelihood of reducing overall DOD injury rates. These activities started a collaborative effort among injury specialty programs at the center and the Defense Medical Surveillance System (DMSS), one of the Military Health System (MHS) Executive Information Decision Support Systems.
In November 2002 a group that had preceded the working group published a report outlining a surveillance process for military injuries based largely on concepts presented in Figure 1 [PDF]. The report proposed that DMSS be the military injury data repository and pointed out advanced health-surveillance capabilities that included multi-source data fusion for epidemiological analysis of the highest quality.
By late 2002 a tri-service collaborative group of audiologists had established a set of ICD-9/CPT standard coding guidelines for military audiology/hearing conservation. The audiology group's intention was to standardize the AcI outcomes data on acoustic injury across the MHS. The ICD-9 data on acoustic injury would be available in the DMSS for surveillance purposes.
Early Acoustic Injury Surveillance
The basic tenets of multidisciplinary surveillance methods for acoustic injury were established in the early 1990s. Hearing conservation audiologists at the U.S. Army Environmental Hygiene Agency (the precursor to the CHPPM) collaborated with epidemiologists and statisticians at the Preventive Medicine department at the Uniformed Services University of Health Sciences and the Medical College of Virginia.
Their studies reviewed a proposed draft American National Standards Institute standard on audiometric database analysis for evaluating hearing conservation programs' performance. The draft standard methods were evaluated against standard public health data analysis processes. The evaluation concluded that the public health outcomes analysis methods for evaluating the performance of hearing conservation programs were superior to the methods proposed in the draft standard.
The collaboration to evaluate the effectiveness of hearing conservation programs continued to 2000. By then the Army had established its Center for Health Promotion and Preventive Medicine and the center added epidemiology, disease surveillance, and other new directorates.
The center's organization now had the appropriate knowledge and skill sets from subject matter experts to continue to develop effectiveness metrics for hearing conservation programs on its own in collaboration with other injury specialties. Additional data sets in the form of ICD-9CM clinical outcomes for different injury specialties became available for analysis in addition to the hearing-conservation program audiometric data.
Postdeployment Surveillance in Wartime
Further evaluation of acoustic injury came in 2004. The standardization of AcI outcomes from MHS ICD-9CM data offered an opportunity to evaluate acoustic injury due to combat exposures during Operation Enduring Freedom (OEF) in Afghanistan and Operation Iraqi Freedom (OIF) deployments. The initial analysis encompassed a one-year period of troops participating in Iraq combat operations and redeploying to home bases, where they received postdeployment medical evaluations that included audiology exams.
We had developed an acoustic-injury ICD-9CM list for analysis of the codes most closely associated with noise-induced hearing loss in a population of soldiers based in garrisons. The acoustic injury code list comprised approximately 18 ICD-9 codes, including acoustic trauma, noise-induced hearing loss, tinnitus, hearing loss profiles, and eardrum perforations (the only other outcome associated specifically with blast exposure from combat).
For this initial study only audiology data were analyzed. Deployment status was defined by an ICD-9CM code used within the MHS to designate postdeployment encounters. Analysis of these data in 2004 and 2005 revealed the postdeployment cohort at significantly higher risk for acoustic injury than the nondeployment cohort. We pointed out the limitations of our initial exploratory study, along with the refinements to be made in later studies, to improve the quality of our analyses and information.
In 2006 we were advised of ICD-9 codes for dizziness and imbalance disorders associated with blast trauma and included them in our 2006 analysis. We expanded evaluation of the acoustic-injury and dizziness/imbalance codes to include soldier data from all clinics, inpatient and outpatient encounters within MHS, and purchased care. This expansion would capture truer prevalence of acoustic injuries than our initial study.
We also sought to verify deployment status by integrating our medical record data with demographic data (including deployment time) from personnel databases. This integration was our first application of multi-source data fusion and analysis.
Our 2006 study encompassed a three-year period covering OEF/OIF deployments. Integrating medical and personnel data showed that the postdeployment ICD-9 designator code was 88.5% accurate when applied, but it wasn't applied in a large number of actual postdeployment encounter cases. The point taken was that deployment status had to be verified in future studies from independent personnel data sources.
The three-year study revealed that the rates of acoustic injury continued to be higher for the postdeployment cohort. The acoustic-injury and dizziness/imbalance codes indicated surges associated with postdeployment cycles for the first two years studied.
Our first results did not include evaluation for traumatic brain injury (TBI), but we realized that TBI would probably be a significant comorbidity to acoustic injury and that acoustic injury, therefore, might be a sentinel diagnosis for TBI. We further realized that this would increase the complexity of our analyses and require a higher level of effort.
In 2007 we began to plan a five-year study that would add TBI and speech-language disorders and other blast trauma central nervous system disorder comorbidities and sequelae to our analysis schema. At this time, polytrauma and mild TBI from blast exposure were becoming more recognized among postdeployment health issues. To evaluate the additional medical outcomes associated with blast trauma, we solicited ICD-9 data categories from other clinical specialties working with concussion and polytrauma patients.
The acoustic-injury/blast trauma ICD-9 code inventory was expanded to include more than 75 codes in categories of acoustic injury, dizziness/imbalance disorders, concussion, and central nervous system comorbidities/sequelae to TBI.
Expanding the ICD-9 codes followed the emerging consensus that blast exposures from combat operations require a multidisciplinary approach to evaluation and treatment (see The ASHA Leader, July 11, 2006, for articles about the multidisciplinary approach to clinical evaluation and treatment of blast injuries.) A preliminary list of providers involved in polytrauma evaluation and treatment is displayed in Figure 2 [PDF] on p. 17.
Hoge et al. (2008) summarized the need for a multidisciplinary clinical approach, noting that the strong associations between mild TBI, post-traumatic stress disorder, depression, and physical health symptoms in combat veterans reinforce the need for a multidisciplinary approach centered in primary care. Evidence-based studies of the management of symptom-based disorders and collaborative care approaches to the evaluation and treatment of coexisting mental disorders in primary care settings are important in designing intervention strategies.
Prevention and Treatment
Over the past 15 years we have come a long way from acoustic injury surveillance for evaluating prevention effectiveness. We have also come to see the importance of gathering information on the rates of additional types of injury and disability among warfighter soldiers. For an army at war, practice-based evidence from public health data has potential to inform clinical practice in terms of the types of health care resources needed for the long-term care and treatment of injured combat veterans.
This information also has potential for providing a basis for searching out best practices and moving toward evidence-based practice and evidence-based clinical practice guidelines. We are becoming a learning community of public health and clinical practice that is connecting, relating, communicating, adapting, and growing to best serve soldiers who have served and sacrificed with honor.
The author acknowledges the significant contributions of USACHPPM colleagues including biostatistician Robyn Lee and epidemiologists Nikki Jordan and Paul Pietrusiak for their work on the AcI/TBI surveillance methods development project.