Respiratory Disease in Army Recruits: Surveillance Program Overview, 1995–2006

The ARDS Program Setting and Surveillance Population 

The ARDS program is conducted at U.S. Army initial entry training (IET) sites. IET for trainees consists of two parts: basic combat training (BCT), a 9-week course in basic soldiering skills, and advanced individual training (AIT), an occupation-specific training lasting up to 2 years following BCT. Contingent on occupation requirements, the BCT and AIT phases of IET may be combined (one-station unit training [OSUT]), in which the cohort remains intact for basic and occupational training. There are currently five installations in the U.S. where BCT and OSUT occur: Ft. Knox KY, Ft. Jackson SC, Ft. Leonard Wood MO, Ft. Sill OK, and Ft. Benning GA.

Within the first 3 days of entering training, trainees are administered vaccinations for several conditions, including those responsible for high ARD rates, as part of medical in-processing. These include vaccinations against measles; mumps; rubella; tetanus; diphtheria; pertussis (since 2006); hepatitis A (since 1995); hepatitis B (since 2002); varicella (since 1999); meningococcus; adenovirus 4, 7 (since 1971; vaccine production ceased in 1996; complete depletion of stocks in 1999); and influenza (seasonal).19 In addition, Ft. Leonard Wood, Ft. Sill, and Ft. Benning administer one dose of 1.2 million units BPG IM as a tandem chemoprophylaxis against GAS/ARF and as a means of reducing the prevalence of streptococcus.20

Components of ARDS Reporting 

Three key components define the Army ARDS program: active clinical reporting of ARD cases, GAS testing of ARD cases, and passive-provider reporting of sentinel cases. The ARDS program defines a clinical case of ARD as a trainee who has an oral temperature ≥38°C with symptoms consistent with a recent onset of acute respiratory infection and who was given job restrictions (i.e., limited-duty status) by a healthcare provider. Clinical case reports are typically received from three medical treatment facilities per participating installation: the troop medical clinic, the hospital emergency department, and the inpatient wards. A rapid strep test (since 2000) or a throat culture, or both, are performed on >80% of ARD cases.

The active clinical reporting component of the Army ARDS program entails installation ARD case reporting to installation public health assets, consolidation with installation at risk population numbers, and subsequent dispatch to the Army Medical Surveillance Activity (AMSA) for analysis. Installation reports include a list of cases and corresponding GAS results; a description of trainees at risk, including unit identification and type and week of training; and type of housing. Following analysis and compilation at AMSA, results (e.g., trend analysis) are returned to participating installations on the same day received.

The passive component of the Army ARDS program involves monitoring the Reportable Medical Events System (RMES), the U.S. Army's official reporting system for conditions of public health concern, for sentinel events of interest.21 The RMES requires reporting of events listed by the U.S. CDC as Nationally Notifiable Infectious Diseases in addition to several military-specific threats.22, 23 With regard to ARD and GAS, the three diseases within the RMES of most interest to this review are ARF, invasive group A streptococcus (invasive GAS), and streptococcal toxic shock syndrome (STSS). The three ARD- and GAS-relevant diagnoses, in addition to other respiratory diseases (e.g., meningococcal disease, tuberculosis) reported to RMES, prompt the interventions defined in the ARDS program guidance.19

Metrics for the ARDS Program 

According to the ARDS program guidance and for the purpose of this review19:

The ARD rate characterizes the number of incident cases among the at-risk trainee population by week. The strep recovery rate indicates the percentage of throat culture or rapid strep tests which are positive. The SASI, or strep-ARD surveillance index, combines both the ARD rate and the strep recovery rate into a single metric. It has been shown to be a sensitive indicator of emerging outbreaks based on past Army experience.18 Established thresholds for intervention, based on clinical experience and expertise, have been established at 1.5% for the ARD rate and and 25.0 for SASI. Exceedance of either threshold for 2 consecutive weeks is defined as a potential outbreak and prompts investigation at the installation level.

Data Analysis 

The Defense Medical Surveillance System (DMSS) is the central repository for all service members' medical encounter and demographic data.24 Trainee rosters for each installation, according to the U.S. Army Training and Doctrine Command (TRADOC), have been available in DMSS only since January 1, 2001. Trainee demographic data from 1995–2000 were unavailable to reflect the early years of this surveillance overview. Analyses were performed in 2007 using SAS version 9.01.

This study is an update of an earlier analysis of the ARDS program from 1985 to 1994. It does not include surveillance data from Ft. Dix NJ or Ft. McClellan AL, because these installations stopped BCT training in 1992 and 1999, respectively.18

Demographics 

Ft. Jackson had the greatest number of trainees, 174,951 over the 6-year period from 2001–2006, followed closely by Ft. Benning and Ft. Leonard Wood (Table 1). Trainees were primarily men. Ft. Knox, Ft. Sill, and Ft. Benning had <1% female trainees, while Ft. Leonard Wood and Ft. Jackson had 37% and 28% female trainees, respectively. Whites represented the highest proportion of trainees at each installation. The distribution of education level was similar across installations, with >63% of trainees possessing a high school diploma but <4% with a college degree. It is important to note that information on education level was less complete than other demographic characteristics.

Table 1. Characteristics of trainees at each initial entry training (IET) installation (2001–2006)a

		Ft. Knox		Ft. Jackson		Ft. Leonard Wood		Ft. Sill		Ft. Benning
		No.	%	No.	%	No.	%	No.	%	No.	%
	TOTAL RECRUITS	57,110		174,951		103,568		55,761		124,794
	Gender
	Male	57,086	100.0	110,865	63.4	74,442	71.9	55,693	99.9	124,625	99.9
	Female	24	0.0	64,086	36.6	29,126	28.1	68	0.1	169	0.1
	Age (years)
	17–19	30,624	53.6	92,344	52.8	59,551	57.5	28,923	51.9	66,160	53.0
	20–24	19,987	35.0	59,245	33.9	32,812	31.7	19,708	35.3	44,339	35.5
	25–29	4,627	8.1	16,093	9.2	7,813	7.5	4,940	8.9	10,597	8.5
	30–35	1,872	3.3	7,269	4.2	3,392	3.3	2,190	3.9	3,698	3.0
	Race/ethnicity
	White, non-Hispanic	41,181	72.1	105,726	60.4	75,012	72.4	41,455	74.3	93,935	75.3
	Black	7,222	12.7	38,383	21.9	13,837	13.4	6,315	11.3	11,392	9.1
	Hispanic	6,065	10.6	20,753	11.9	9,959	9.6	5,381	9.7	12,886	10.3
	Other/Unknown	2,642	4.6	10,089	5.8	4,760	4.6	2,610	4.7	6,581	5.3
	Education level
	<High school	6,568	11.5	21,420	12.2	18,425	17.8	9,134	16.4	15,178	12.2
	High school	36,992	64.8	105,908	60.5	59,248	57.2	33,552	60.2	78,506	62.9
	Some college	2,032	3.6	6,537	3.7	3,890	3.8	2,067	3.7	4,556	3.7
	College degree	1,014	1.8	5,542	3.2	2,806	2.7	1,688	3.0	3,572	2.9
	Unknown	10,504	18.4	35,544	20.3	19,199	18.5	9,320	16.7	22,982	18.4
	Home of residenceb
	South	22,057	38.6	75,848	43.4	39,174	37.8	21,799	39.1	47,258	37.9
	Midwest	13,621	23.9	38,360	21.9	27,671	26.7	14,275	25.6	30,272	24.3
	West	11,875	20.8	32,782	18.7	19,743	19.1	10,974	19.7	26,216	21.0
	Northeast	8,297	14.5	24,193	13.8	15,254	14.7	7,769	13.9	17,998	14.4
	Unknown	1,260	2.2	3,768	2.2	1,726	1.7	944	1.7	3,050	2.4
	Service component
	Active	42,317	74.1	128,105	73.2	69,467	67.1	38,345	68.8	94,838	76.0
	Guard	14,078	24.7	42,136	24.1	30,700	29.6	16,436	29.5	29,064	23.3
	Reserve	715	1.3	4,710	2.7	3,401	3.3	980	1.8	892	0.7

IET, initial entry training

a U.S. Army Training and Doctrine Command (TRADOC) rosters for each installation have been available in the Defense Medical Surveillance System (DMSS) since 2001. b Region division based on U.S. Census Bureau and Office of Management and Budget classification.

Rates of ARD 

A total of 96,789 ARD cases were identified between 1995 and 2006 (Table 2). The median weekly ARD rate was greater among women than men (0.66% vs. 0.54%) for all years except 2001, 2002, and 2003. Average weekly ARD rates were consistently higher from 2000 to 2006 than from 1995 to1999. In addition, the rate ratios of winter rates to summer rates were greater than 1.0 prior to 2000, but from 2000 to 2006, ratios were consistently near 1.0 (Figure 1).

Table 2. Yearly summary of ARDS indicators (January 1995–December 2006)

		Trainee-weeks		ARD cases (rate/100/gender)		ARD ratea(%)	No. of GAS tests performed (% of year total ARD cases)	No. of GAS (+) results	Strep recovery rateb(%)	No. of RMES casesc
	Year	Men	Women	Men	Women
	1995	899,855	145,115	1,898(0.21)	668(0.46)	0.25	2,105(82.0)	162	7.69	0
	1996	996,546	203,622	1,391(0.14)	802(0.39)	0.18	1,650(75.2)	144	8.73	0
	1997	1,007,202	229,983	2,239(0.22)	813(0.35)	0.25	3,052(100.0)	227	5.43	0
	1998	1,028,670	202,868	2,735(0.26)	949(0.47)	0.30	3,269(88.7)	170	5.20	1
	1999	1,144,596	282,394	4,805(0.42)	2,226(0.79)	0.49	7,031(100.0)	302	3.76	1
	2000	1,290,170	331,877	8,866(0.69)	2,372(0.71)	0.69	11,238(100.0)	367	3.02	0
	2001	1,349,063	315,955	12,582(0.93)	2,432(0.77)	0.90	14,766(98.3)	633	4.29	2
	2002	1,370,476	295,344	9,663(0.70)	1,948(0.66)	0.70	10,436(89.9)	1,024	9.81	2
	2003	1,288,236	298,764	10,986(0.85)	2,302(0.77)	0.84	12,044(90.6)	359	2.98	5
	2004	1,311,052	272,574	7,655(0.58)	2,529(0.93)	0.64	9,485(93.1)	257	2.71	0
	2005	1,156,250	205,210	6,592(0.57)	1,559(0.76)	0.60	7,182(88.1)	580	8.08	2
	2006	1,396,989	271,749	7,188(0.51)	1,589(0.58)	0.53	8,162(93.0)	892	10.93	8
	Total	14,239,105	3,055,455	76,600(0.54)	20,189(0.66)	0.57f	93,447(96.5)	5,117	5.32d	21

ARD, acute respiratory disease; ARDS, acute respiratory disease surveillance; GAS, group A streptococcus; No., number; RMES, reportable medical events system.

a Average weekly ARD rate (per 100 trainee-week) by year b Strep recovery rate=(Total Throat culture (+) or Rapid strep (+)/Total test performed) × 100 c Includes only invasive GAS, toxic shock syndrome, acute rheumatic fever d Median

View Large Image Download to PowerPoint

Figure 1. ARD rate and rate ratios by year and season.

Differences in ARD rates and GAS rates were seen across installations. The median ARD rates for 1995–2006 were 0.59%, 0.56%, 0.36%, 0.08%, and 0.07% for Forts Jackson, Leonard Wood, Benning, Knox, and Sill, respectively. Yearly rates were consistently highest at Ft. Leonard Wood and lowest at Ft. Knox.

Outbreaks, as defined by the ARDS program guidance threshold criteria, also varied by installation. Ft. Knox experienced the least number of outbreaks, only one during summer 2003, and Ft. Leonard Wood the highest, at nine outbreaks occurring mostly in the winter season. No ARD outbreaks were reported between 1995 and 1998.

Strep Recovery Rates 

On an annual basis, more than 80% of ARD cases were tested by either a GAS rapid strep test or a throat culture. This proportion varied by installation (data not shown). Strep recovery rates peaked in 1995, 1996, 2002, 2005, and 2006, although these peaks did not correspond with ARD peaks (Table 2). On average, 4% of ARD cases tested were positive for GAS, excluding the aforementioned peak years for strep recovery rates.

Sentinel Cases 

Twenty-one sentinel case reports were submitted via the passive system from 1995 through 2006 (Table 2). Invasive GAS was the most common GAS-related RMES sentinel diagnosis during the 12-year period. The high number of invasive GAS cases observed in 2003 and 2006 were attributable to GAS outbreaks at Ft. Knox and Ft. Leonard Wood, respectively. No cases of STSS were reported in this period, and only one ARF case (1998, Ft. Leonard Wood) was reported in this period.

Outbreaks of GAS (+) 

Five selected GAS outbreaks are highlighted in this analysis, four of which were investigated and interventions implemented by installation public health. The fifth outbreak, at Ft. Leonard Wood in 2006, prompted a formal investigation by the U.S. Army Medical Command.

In December 2002, Ft. Sill reported a sentinel case of invasive GAS (pneumonia). Strep recovery rates during this outbreak were as high as 89%. Targeted mass BPG chemoprophylaxis (in addition to the regular administration of tandem BPG prophylaxis during medical in-processing) to high-risk units (i.e., high ARD counts in specific units found by information in the clinical reporting component of the ARDS program) was implemented to prevent further GAS-related illnesses. However, ARD thresholds were never exceeded for 2 consecutive weeks. In contrast, the SASI exceeded the threshold for more than 2 consecutive weeks in February 2002 and sporadically throughout the winter season. There were no reports of BPG shortages during this outbreak.

Ft. Knox experienced an outbreak in the summer and early fall of 2003. A sentinel event, a fatal case of invasive GAS (necrotizing fasciitis), occurred in June 2003, followed by four more nonfatal cases (one of necrotizing fasciitis, one of prepatellar bursitis, and two of cellulitis) within 3 months. Only the ARD rates reached outbreak levels in late August. During 2 of the 4 weeks, GAS was isolated in 30%–40% of ARD cases tested. Implementation of enhanced preventive measures included thorough cleaning of shared equipment, frequent cleaning of canteens, increased availability of soap, reorganization of barracks space, and basic hygiene education. The invasive cases subsided without the administration of BPG during this outbreak.

An additional invasive GAS outbreak occurred at Ft. Knox in 2006. Neither the ARD nor SASI threshold was exceeded, nor had sentinel reports been sent to RMES. However, preventive medicine personnel were alerted to the possibility of a virulent strain of GAS by a trainee who had been admitted to the hospital for GAS (+) pneumonia. Within 2 weeks, two additional cases from the same training unit were admitted to the hospital. In response, BPG-targeted prophylaxis was administered to all trainees in that unit.

The last two highlighted outbreaks occurred at Ft. Leonard Wood. The first outbreak presented with a sentinel case (STSS) in December 2005, and both the ARD and SASI indices exceeded outbreak levels. Three additional trainees were admitted for GAS (+) pneumonia. In response, targeted mass BPG prophylaxis was initiated. The decision for targeted mass chemoprophylaxis versus installation-wide (trainee) mass chemoprophylaxis was influenced by the need to maintain an emergency stockpile. Of note was a shortfall in the BPG supply for the provision of chemoprophylaxis to incoming trainees that had been reported since September 2005. Following the return of trainees from holiday leave, a high SASI continued for an additional 5 weeks until the beginning of spring.

In the fall of 2006, seven sentinel cases of GAS (one of STSS, two of necrotizing fasciitis, and four of pneumonia) were reported from Ft. Leonard Wood. In response, only units with high ARD and GAS (+) counts were targeted for BPG prophylaxis. However, new GAS (+) cases and high counts of ARD were then observed in other units. SASI levels exceeded thresholds sporadically in October and November, but never for 2 consecutive weeks. A policy of mass prophylaxis with BPG and tandem prophylaxis with oral penicillin VK (due to a shortage of BPG), in addition to basic preventive medicine measures, was instituted in response to the continued high rates of disease, and ARDS indices returned to normal.

Group A streptococcal M-typing during these outbreaks varied by site. During the Ft. Sill outbreak, no M-typing was done, and data from the 2003 Ft. Knox outbreak were not available. During the 2005 Ft. Leonard Wood outbreak, all ten isolates tested were M5. During the 2006 Ft. Leonard Wood outbreak, 31 isolates were tested: 24 were M5; three were M101; two were M77; and two were M18. Thirty isolates were tested during the 2006 Ft. Knox outbreak, and M5 was identified in 23, M4 in six, and M2 in one isolate. Only the M18 cases exhibited mucoid morphology.25