INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

As rates of tuberculosis in the United States decline, the disease will increasingly become concentrated in populations of high-risk and difficult-to-access patients. Among these are individuals who are confined in prisons and jails. Incarcerated populations have high rates of substance abuse, human immunodeficiency virus (HIV) infection, latent tuberculosis infection, low socioeconomic status, and other risk factors associated with active tuberculosis (1, 2). In Memphis, 43% of persons reported to have tuberculosis have been incarcerated in that city's jail (3), 20% of tuberculosis patients in Chicago are released inmates of the Cook County Jail (4), and 24% of tuberculosis cases in Nassau County, NY were associated with the jail there (5). Transmission of tuberculosis from prisons into surrounding communities has been documented (6), and correctional facilities now are considered important reservoirs of infection (7). Although the rate of tuberculosis in the general population in the United States in 1998 was 6.8 cases per 100,000 (10), many jail and prison systems have reported rates of 200 per 100,000 persons and higher (3, 6, 11).

Given the high-risk population in jails and the substantial potential for transmission of tuberculosis both within and out of these facilities, appropriate screening for active tuberculosis and management of disease in jails is critical. Most jails and prisons perform tuberculin skin testing (TST) to screen for tuberculosis (19), though this method is neither sensitive nor specific for active disease (20) and is very difficult to complete successfully in jails (3, 4, 18). Routine roentgenographic screening for tuberculosis in the general population in the United States is not recommended. In recent years, however, it has been suggested that screening jail inmates for active tuberculosis with chest radiographs may be appropriate (4, 17, 18, 20, 21). The availability of new high-speed, low-dose miniature chest radiograph technology makes this a safer and more practical option than in the past. Several jail systems are now using universal chest radiograph screening for all admissions (22). We evaluated the cost effectiveness of miniature chest radiograph screening for tuberculosis on admission to jail, under varied conditions.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cost-effectiveness analyses were performed using Decision Analysis by TreeAge (DATA 3.5) software (23). Several methods of screening for active tuberculosis on admission to jail were evaluated, including routine miniature chest radiography, tuberculin skin testing, and symptom screening.

Rates of tuberculosis, prevalence of HIV infection, and values of other variables associated with screening were based on estimates in the published literature (Table 1) (1, 4, 10, 14, 17, 18, 21, 24). Sensitivity analyses were performed using a wide range of values for many of these parameters, derived from studies conducted under a variety of conditions (see online data supplement). A 3% discount rate was used for future costs and outcomes.

The baseline incidence of active tuberculosis in a high-risk jail population was estimated to be 68 per 100,000 (4). Calculations were also performed using rates from other studies in jails (4, 9, 18), as well as the rate in the general U.S. population (10). We estimated that each case of active tuberculosis detected through screening at admission would result in half the number of infected contacts and future cases of active disease compared with a passively diagnosed case. Of persons with tuberculosis, 1.1% are assumed to be infected with a multidrug-resistant (MDR) organism (10). The prevalence of HIV infection in the jail population was estimated at 0.5% (17).

We estimated that 25% of jail inmates have a positive tuberculin skin test (TST) (46). The sensitivity and specificity of various screening methods were based on the published literature, using estimates from studies specifically in jails when available (Table 1). Chest radiographs were predicted to be suggestive of possible tuberculosis in 5% of jail inmates. Screening for symptoms of active tuberculosis with a questionnaire was estimated to lead to a recommendation for further evaluation in 5.6% of inmates. We assumed that an average of 150 persons would be screened daily (3).

Costs were adjusted to 1998 dollars. Baseline analyses were performed assuming that a case of active tuberculosis costs an average of $16,640 to treat (including both inpatient and outpatient therapy) (28). Treatment of multidrug-resistant tuberculosis was estimated to cost $230,000 (30). We estimated that infected contacts of cases without active disease would receive preventive therapy at a cost of $229 (27).

Miniature chest radiograph screening was estimated to cost $6.60 per person screened (based on equipment costs of $350,000 and facilities modifications of $100,000, amortized over 10 yr; $18,000/yr in service costs after a 2-yr warranty; $0.26/person in supplies; 24-h technician staffing at $20.00/h; and radiologist services at $110,000/yr, assuming 150 persons screened per day). Tuberculin skin testing was estimated to cost $8.00 per person screened. Screening by symptom questionnaire was estimated to cost $4.80 per person screened (including full-time staff coverage at $30.00/h). The cost of evaluating a suspect for active tuberculosis after a positive screening test was estimated to be $180 (including initial and follow-up medical evaluations at $94 and $40, respectively (53), and chest radiograph at $46).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

At baseline assumptions, we calculated an average total cost of $6.60 per inmate for routine radiograph screening at admission. The cost of screening for active tuberculosis with miniature chest radiography was determined to be $9,600 per case identified, compared with $32,100 per case identified through tuberculin skin testing, and $54,100 per case with symptom screening. In addition, screening with miniature chest radiography was estimated to identify 0.68 cases per 1,000 inmates screened, compared with 0.25 cases with tuberculin screening and 0.09 cases with symptom screening.

In a high-risk and high-volume setting, the total cost of miniature radiographic screening, evaluation, and treatment per case of active tuberculosis identified would be $37,400. Total costs associated with tuberculin skin testing or symptom questionnaire were estimated to be $60,300 and $84,100 per active case identified, respectively.

Assuming a lower estimate of 0.58 for the specificity of miniature radiograph screening for active tuberculosis, screening in this manner would cost $46,600 per active case of tuberculosis identified (Table 2). Even under this adverse assumption, miniature chest radiography would remain more cost effective than symptom screening.

Under baseline assumptions, increasing the estimated cost of miniature radiograph screening to $25 per person changed the estimate of the cost of screening to $36,500 per active case of tuberculosis detected. Varying the probability of a positive chest radiograph on initial screening from the minimum to maximum estimates changed the estimate of the cost of screening to $21,500 and $5,700 per case, respectively. If the prevalence of tuberculosis among the inmate population was estimated to be similar to that of the general U.S. population (6.8/ 100,000), miniature radiographic screening was estimated to cost $62,500 per case identified.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The cost of screening newly admitted jail inmates using miniature chest radiography was estimated at approximately $9,600 per case of active tuberculosis identified. These calculations were based on reasonable baseline assumptions, using estimates of the costs and probabilities associated with various methods of screening for active tuberculosis derived from studies in correctional facilities and similar settings. This compares favorably with the cost effectiveness of many other widely recommended public health preventive interventions, and represents less than half of the cost of treating a single case of disease due to a drug-susceptible organism (30, 31). Even as the estimated incidence of active tuberculosis in the jail population was estimated to decline, screening with routine miniature chest radiography remained cost effective. If the incidence of tuberculosis in this population reflected the much lower incidence in the general U.S. population, miniature radiographic screening was estimated to cost only $62,500 per active case identified.

Screening for active tuberculosis with miniature chest radiography was determined to be more sensitive and more cost effective than screening with either tuberculin testing or symptom screening. In the setting of a high-turnover, chaotic urban jail, there is substantial evidence that both of the latter strategies are inadequate methods for screening. In the Memphis jail, a well-intentioned program to screen inmates with tuberculin skin testing on admission resulted in only 5.8% of admitted inmates actually having skin tests read successfully (3). Other jails have had similar problems (18, 34), and in the Cook County jail 43% of inmates noted to have a positive TST were discharged before their follow-up evaluation was complete (4). In the latter study, the mean time from admission to jail until isolation of tuberculosis suspects was 17.6 d with TST screening versus 2.3 d with miniature radiograph screening (4). The Cook County jail doubled the number of cases of active tuberculosis detected when radiographic screening was introduced there (4).

Even under ideal circumstances, tuberculin testing is neither sensitive nor specific for active tuberculosis (20). In one study of a high-risk population it was estimated that 65% of persons with radiographs most suggestive of tuberculosis would have been missed with a TST screening program (34). Tuberculin testing clearly has an important role in screening for infection and evaluating persons suspected of having active tuberculosis. Its limitations, however, are compounded in settings such as jails, where high turnover and limited follow-up make it unsatisfactory as a primary method of screening for active disease. Regardless of the cost estimates of radiography or other alternatives, reliance solely on tuberculin skin testing is inadequate for the detection of active tuberculosis in jails.

Similarly, relying on symptom screening for detection of active tuberculosis has many drawbacks in jails. This method of screening is appealing because it is inexpensive, requires no specialized equipment, and is directed at identifying persons most likely to be infectious to others. Nonetheless, its already limited sensitivity and specificity (21, 37, 40, 50, 51, 54), are likely worsened in a setting where many persons being screened may be under the influence of drugs or alcohol, have mental illness, or have substantial incentives for either over- or underreporting symptoms (21).

The use of chest radiography to screen for tuberculosis has not been widely advocated in the United State for decades (18). The declining incidence of disease, expense of the technique, and concerns about the risks of radiation contributed to its demise as a screening tool for the general population. Recent calls for renewed efforts to target screening to high-risk populations, with the goal of eliminating the disease in this country, necessitate reevaluation of all available tools for combating tuberculosis (55). Recent technological advances have led to the availability of high-quality miniature chest radiographs that can be performed in under 60 s, and at approximately one-tenth of the radiation dose of a conventional radiograph (4). Taking multiple images on spooled film or use of digital technology have the added advantages of increased speed, easy transportability and storage, and decreased associated costs. These improvements in technology make chest radiography particularly attractive for use in screening of high-risk populations such as jail inmates. Chest radiography is more sensitive and specific for active disease than other available screening methods, is less dependent on patient cooperation than symptom screening, provides objective data for decision making, and allows completion of screening within hours rather than the days required with tuberculin testing. In one large urban jail, institution of miniature radiographic screening doubled the number of cases identified (4). As elimination of endemic tuberculosis in the United States becomes a realistic possibility, priority should be given to implementing the most sensitive practical methods of screening, particularly focused on the highest-risk populations.

Chest radiography has the relative disadvantage of not identifying most persons with tuberculosis infection but without active disease. Under many circumstances such persons would be candidates for evaluation for treatment of latent tuberculosis infection (LTBI). In jails, however, the majority of persons are incarcerated for a day or less. Data suggest that attempts to complete preventive therapy for tuberculosis after discharge from correctional facilities are abysmally unsuccessful (46). Under such circumstances, identification of active disease and prevention of spread should be a priority, and radiographic screening would meet that need pragmatically in such settings.

The high "up front" cost of purchasing and installing rapid miniature chest radiographic equipment is a substantial barrier to implementing the technology in many settings. Despite arguments that over time the per unit cost of screening with radiography actually is less than that of other methods, initial allocation of several hundred thousand dollars for the equipment may make this method of screening impossible for many correctional systems. The primary mandate of correctional facilities is securely incarcerating inmates, and any demands to participate in public health interventions that might impede this priority are likely to be met with resistance. The public health sector will need to strongly encourage and support the criminal justice system to foster the cooperation necessary to implement this type of screening.

Jails and prisons are an underappreciated "golden opportunity" to reach some of the highest-risk and most difficult-to-reach populations to detect and prevent tuberculosis. Over 10 million persons pass through U.S. correctional facilities each year, and 5 of 1,000 Americans are behind bars on any given day (56). Over 130,000 persons with tuberculosis infection are released to the community each year from state and federal correctional facilities alone (19). In some large cities, 20% to 46% of persons with active tuberculosis are ex-inmates of the local jail (3, 4). As rates of tuberculosis continue to fall, correctional facilities are likely to remain one of the last strongholds of endemic disease in this country. Although the cost effectiveness of any method of screening for disease will decline as we get closer to the goal of elimination, we must be willing to aggressively use all tools at our disposal if the goal is to be met. The public health community must be willing to work closely with correctional facilities to take advantage of the opportunity to intervene in an important public health problem while minimizing disruption of the facilities' law-enforcement functions. Rapid miniature chest radiography on admission to jail should be evaluated carefully as an effective intervention in high-risk facilities.