Evaluation of tuberculosis contact investigations in California

Transcription

1 INT J TUBERC LUNG DIS 7(12):S363 S IUATLD Evaluation of tuberculosis contact investigations in California J. E. Sprinson, J. Flood, C. S. Fan, T. A. Shaw, L. Pascopella, J. A. Young, S. E. Royce California Department of Health Services, Tuberculosis Control Branch, Berkeley, California, USA SUMMARY OBJECTIVE: To estimate the burden of tuberculosis (TB) contact investigations in California, assess outcomes and effectiveness, and identify performance gaps. METHODS: Aggregate program management reports were used to examine contact investigations conducted for pulmonary TB cases reported between 1 July 1999 and 30 June 2000 in California. Findings were compared to national objectives, and performance gaps were identified. Costs were estimated, and effectiveness of TB case detection and prevention was assessed. RESULTS: A total of 2032 acid-fast bacilli sputum smearpositive and sputum culture-positive cases was reported; contacts were elicited, and (88) contacts were evaluated. TB disease and latent tuberculosis infection (LTBI) were diagnosed in 111 ( 1) and 4609 (30) contacts, respectively; 1958 (43) contacts with LTBI completed treatment. Costs of contact investigations were estimated at $4.8 million; 81 of expected TB cases were detected, but only 35 of cases expected to occur within 2 years following the investigation were prevented. CONCLUSIONS: California s performance did not meet national objectives for contact evaluation or treatment completion; improved effectiveness of contact investigations in California is needed. Although analysis of existing contact investigation surveillance data provided a macro-level view of performance gaps, expanded surveillance data are required to inform interventions. KEY WORDS: tuberculosis; contact tracing; surveillance; communicable diseases; outcome assessment; program evaluation WITHIN THE United States, tuberculosis (TB) impacts the state of California disproportionately. With approximately 12 of the US population, California has contributed approximately 20 of the nation s incident cases annually for many years. 1,2 In 2001, the state reported 3332 cases, the first year without a decline in case numbers since 1993, with a case rate of 9.5 per population, 2 the second highest nationally. 1,2 In the US, contact investigations are conducted when suspected or confirmed pulmonary or laryngeal TB cases are reported. Investigations include: case interviews to elicit contacts; testing and evaluation of contacts to identify individuals with latent tuberculosis infection (LTBI) and TB disease; and treatment for TB disease, or for LTBI to prevent progression to TB disease. Contacts are typically prioritized according to the infectiousness of the associated case. Given the magnitude of California s case load, there is a large work burden associated with contact investigation activities. Also, it is likely that California s contact investigation performance substantially influences national performance. TB morbidity is distributed non-uniformly across California s 61 local health jurisdictions. Nineteen jurisdictions accounted for 93 of TB cases in 2001, with case loads ranging from 25 to Eleven (18) jurisdictions reported fewer than five cases, and 13 (21) reported no cases. 2 California s jurisdictions are responsible for reporting and providing oversight for TB cases and conducting contact investigations associated with these cases. The California Department of Health Services Tuberculosis Control Branch (TBCB) allocates state and federal funds to jurisdictions for these activities. Staffing and resources vary widely across jurisdictions. Clinical care for TB services is provided via public health clinics, the private medical community, and managed care organizations. TBCB collaborated with the California Tuberculosis Controllers Association to develop statewide contact investigation guidelines. 3 The objectives of this study were to quantify the magnitude and costs of contact investigations conducted in California, assess contact investigation outcomes and effectiveness, and identify key performance gaps. Correspondence to: Joan E Sprinson, California Department of Health Services, Tuberculosis Control Branch, 2151 Berkeley Way, Room 608, Berkeley, CA , USA. Tel: ( 1) Fax: ( 1) jsprinso@ dhs.ca.gov

2 S364 The International Journal of Tuberculosis and Lung Disease METHODS In July 1999, California began using the Aggregate Reports for Tuberculosis Program Evaluation (ARPE): Follow-Up and Treatment for Contacts to Tuberculosis Cases developed under the auspices of the Centers for Disease Control and Prevention (CDC). 4 This onepage report aggregates data for nine variables, stratified by three case types: sputum (both culture-positive and -negative, hereinafter referred to as ); sputum culturepositive (hereinafter smear-negative); and other cases. Prior to beginning its use, TBCB provided training to jurisdictions on ARPE completion. In June 2000, a statewide contact investigation workgroup standardized instructions and definitions for ARPE completion and customized the form for use in California. Most jurisdictions completed the ARPE manually by reviewing each contact record. Jurisdictions faxed or mailed ARPEs to TBCB every 6 months. Contact investigation outcomes for and smear-negative cases were examined using the most complete data reported from two 6-month ARPE cohort periods (1 July June 2000). California s performance for contact elicitation, contact evaluation, and treatment completion for contacts with LTBI was compared to national performance objectives, and the proportion of jurisdictions meeting these objectives was determined. Overall effectiveness of contact investigations in detecting and preventing TB cases was estimated by calculating missed opportunities for case detection and prevention for each contact investigation step and summing this information across all steps. In developing these estimates, cases with no contacts elicited were assumed to have the same number of contacts as the statewide average number of contacts per smearpositive and smear-negative case. Contacts who were not evaluated were assumed to have the same probability of TB disease and LTBI as contacts who completed evaluation. The probability of a contact with LTBI, but without completed treatment, progressing to active TB disease was assumed to be for the first 2 years following diagnosis of LTBI. 5 Costs for contact investigations during the study period were estimated, based on published methods and findings 6 that estimated a cost per contact investigation for drug-sensitive and multidrug-resistant (MDR) TB cases of $2263 and $7329, respectively, in 1998 US dollars. The cost of contact investigations during the study period was estimated by determining the percentage of MDR (1.6) and drug-susceptible (98.4) and smear-negative cases that were counted during this period. These percentages were used with referenced 5 information to estimate the total cost of contact investigations during the study period. Cost estimates were not adjusted for inflation. All analyses were performed using the SAS statistical software package (Release 8.2, SAS Institute Inc., Cary, NC). RESULTS Contact elicitation During the study period, 2032 pulmonary cases and contacts were reported on the ARPE (Table 1); 62 (1267) of the cases were. Three times as many contacts were elicited for these cases as for smear-negative cases. Statewide, a total of 214 (11) cases had no contacts elicited; 101 (8) of cases and 113 (15) of smear-negative cases had no contacts elicited. The mean number of contacts elicited per case was 10.5 (range ) contacts for cases and 5.8 (range ) contacts for smear-negative cases. If three jurisdictions with more than 50 contacts per case were excluded from this calculation, the mean number of contacts elicited per case was 9.1 and 5.0 contacts for cases and smear-negative cases, respectively. Ninety-two per cent of cases had at least one contact identified, exceeding the CDC Table 1 Outcomes of contact investigations: California, 1 July June 2000 Measure n ()* culture-positive n ()* Cases for investigation (62.4) 765 (37.6) Cases with no contacts 101 (47.2) 113 (52.8) 214 Number of contacts (75.1) 4418 (24.9) Mean contacts per case 10.5 (NA) 5.8 (NA) 8.8 Contacts evaluated (76.0) 3744 (24.0) Contacts with TB disease 87 (78.4) 24 (21.6) 111 Contacts with LTBI (80.4) 902 (19.6) Contacts starting treatment (81.1) 575 (18.9) Contacts completing treatment (79.4) 404 (20.6) Source: Aggregate Reports for Tuberculosis Program Evaluation: Follow-Up and Treatment for Contacts to Tuberculosis Cases. * of n/n. NA not applicable; TB tuberculosis; LTBI latent tuberculosis infection. N

3 Evaluation of tuberculosis contact investigations in California S365 Table 2 Performance of contact investigations in California: Centers for Disease Control and Prevention and California objectives, 1 July June 2000 Measure culture-positive CDC objective CA objective Cases with 1 contact elicited * 90* Evaluation rate * 95* TB disease NA NA LTBI NA NA Treatment start rate NA 80 Treatment completion rate Source: Aggregate Reports for Tuberculosis Program Evaluation: Follow-Up and Treatment for Contacts to Tuberculosis Cases. * CDC and CA objectives pertain to contacts of sputum cases. CDC and CA objectives pertain to all contacts with LTBI, regardless of smear status of the associated case. CDC Centers for Disease Control and Prevention; CA California; NA not applicable; LTBI latent tuberculosis infection. objective of 90 (Table 2). Eighty-three per cent (35/ 42) of California s jurisdictions with one or more case met the CDC objective, with performance ranging from 56 to 100 and a mean of 94. Contact evaluation During the study period, (88.6) contacts of cases were evaluated (Table 1); TB disease and LTBI were diagnosed in 87 (0.7) and 3707 (31.3) respectively. For smear-negative cases, 3744 contacts (84.7) were evaluated; TB disease and LTBI were diagnosed in 24 (0.6) and 902 (24.1), respectively. Overall, the state s performance fell short of the CDC objective (Table 2) to evaluate 95 of contacts of cases; however, 31 (13/ 42) of jurisdictions with one or more case met this objective, with performance ranging from 25 to 100 and a mean of 84. Treatment of contacts with LTBI During the study period, 3048 (66.1) contacts with LTBI started treatment, and 1958 (64.2) contacts completed treatment (Table 1). Reasons for stopping treatment were reported for 90.8 (990/1090) of contacts who did not complete therapy: 516 (52.1) contacts chose to stop; 175 (17.7) contacts moved and the follow-up was unknown; 174 (17.6) contacts were lost to follow-up; 79 (8.0) contacts had an adverse effect to medication; 39 (3.9) contacts discontinued treatment based on the provider s decision; four (0.4) contacts developed active TB; and three (0.3) contacts died. Although 31 (14/45) of California s jurisdictions with one or more cases for investigation met the CDC objective of 85 for treatment completion among contacts with LTBI, the state fell short of the objective. Treatment completion by jurisdiction ranged from 17 to 100, with a mean of 72. Estimated effectiveness of contact investigations Contact investigations were estimated to have detected 81.0 (111/137) of the TB cases identifiable through a contact investigation (Table 3). Contact investigations were estimated to have prevented only 34.6 (28/81) of the future TB cases estimated to occur, lacking treatment for LTBI, within 2 years following the investigation (Table 4). Contact investigation cost estimate Costs for contact investigations during the study period were estimated at $4.8 million in 1998 dollars. Information was not available regarding the payer source. Table 3 Estimated effectiveness of contact investigations in detecting TB disease: California, 1 July June 2000 Case detection estimation categories Contacts n culture positive Contacts not elicited Contacts not evaluated Contacts not elicited or not evaluated Contacts not elicited or not evaluated and assumed to have TB disease 18* 8* 26* Contacts diagnosed with TB disease Estimated per cent of TB cases detected Source: Aggregate Reports for Tuberculosis Program Evaluation: Follow-Up and Treatment for Contacts to Tuberculosis Cases. * Estimated number of contacts rounded to nearest integer. TB tuberculosis.

4 S366 The International Journal of Tuberculosis and Lung Disease Table 4 Estimated effectiveness of contact investigations in preventing TB disease: California, 1 July June 2000 Contacts n Case prevention estimation categories culture-positive Contacts not elicited Contacts not evaluated Contacts not elicited or not evaluated Contacts not elicited or not evaluated and assumed to have LTBI Contacts diagnosed with LTBI who did not complete treatment contacts with LTBI without completed treatment* Estimated contacts without completed treatment progressing to TB disease Estimated cases prevented by completed treatment Estimated per cent of TB cases prevented Source: Aggregate Reports for Tuberculosis Program Evaluation (ARPE): Follow-Up and Treatment for Contacts to Tuberculosis Cases. * Includes contacts not elicited or not evaluated and assumed to have LTBI, and contacts diagnosed with LTBI (from ARPE) who did not complete treatment. Probability of a contact with LTBI progressing to active TB disease was assumed to be for the first 2 years following diagnosis of LTBI. Estimated number of contacts rounded to nearest integer. TB tuberculosis; LTBI latent tuberculosis infection. DISCUSSION Our findings underscore the magnitude of California s contact investigations and their potential impact on TB control in the US. With approximately 20 of US TB cases, California likely conducts one in every five contact investigations nationally, with estimated costs in excess of $4.8 million during the study period. Given this burden, special attention should be paid to ensuring that resources are used effectively and efficiently, that contacts at greatest risk of acquiring LTBI and progressing to TB disease are prioritized, and that all steps to detect and prevent disease and interrupt transmission are completed. The first step in the contact investigation is to interview the TB case and elicit contacts. Effective contact elicitation requires experienced interviewers who are culturally and linguistically competent. However, even when these conditions are met, the TB patient may be unwilling or unable to share information about contacts. Three times as many contacts of smearpositive as smear-negative cases were reported. This finding may indicate that jurisdictions prioritized contact tracing among cases presumed to be the most infectious, as the California contact investigation guidelines recommend. The lower proportion of cases without contacts among compared to smearnegative cases is consistent with this finding. Although California exceeded the CDC objective of 90 for contact elicitation, 101 cases and 113 smearnegative cases during the study period had no contacts elicited. Since it is likely that most of these cases exposed other individuals, contacts who were not elicited represent missed opportunities to detect and prevent TB. After contacts are elicited, complete and timely evaluation is needed to detect contacts with TB disease or LTBI. This step, too, may present many challenges. Our findings of newly detected cases per among contacts of cases and newly detected cases per in contacts of smear-negative cases were consistent with previous reports. 7,8 Of note, the disease rate is similar for contacts of and smear-negative cases. This finding may indicate that jurisdictions better prioritized the highest risk contacts of smear-negative cases, or that they cast the net too widely for contacts of cases, or that other, unmeasured factors contributed importantly to similar TB disease rates among contacts of and smearnegative cases. A smaller fraction of contacts of smear-negative cases (24.1) was diagnosed with LTBI than contacts of cases (31.3). This finding may be partially explained by lower transmission risk associated with smear-negative cases, as well as the practice during the study period of at least one large jurisdiction to provide contacts of smearnegative cases with single tuberculin skin tests, even if the result, obtained within 12 weeks of last exposure to the index case, was negative. In order for contact investigations to be effective in preventing future TB cases, contacts with LTBI must start and complete treatment. Our study revealed that only 42.5 of contacts with LTBI both started and completed treatment. A low proportion (66.1) of contacts started treatment for LTBI, and a low proportion (64.2) of those who started, completed treatment. The primary reason why contacts did not complete treatment was reported as contact chose to stop on the ARPE. However, this information leaves crucial questions unanswered. For example, did the contacts stop because they didn t have time or money to return to the clinic, rejected a diagnosis of LTBI because they received BCG as a child, or worried about medication side effects? Additionally, are patients of private providers, in comparison to public providers, or patients without health insurance more likely to stop therapy? The ARPE does not capture information to answer these questions. Reasons why a contact chose to stop treatment may suggest very different interventions to improve treatment outcomes.

5 Evaluation of tuberculosis contact investigations in California S367 The primary objectives of contact investigations are to halt transmission of Mycobacterium tuberculosis by detecting and treating persons with TB disease and to prevent future TB cases by identifying and treating contacts with LTBI. Our analysis of contact investigations conducted in California indicated that 81.0 of the expected number of TB cases was identified. However, contact investigations only prevented 34.6 of expected future TB cases. Prevention effectiveness would likely be even lower if time frames beyond 2 years were considered. These analyses suggested that approximately 20 and 65 of efforts designed to detect and prevent TB cases, respectively, did not fully achieve the desired outcomes. To improve case detection and prevention through evidence-based interventions, collection of additional information locally is needed. For example, reasons why contacts did not start and complete evaluation and treatment are essential. Similarly, information on whether the highest risk contacts (e.g., young children and persons with immunosuppression) were elicited, completed evaluation, and, if infected, completed treatment for LTBI, is critical. Data to support analysis of timeliness of contact investigation steps are also necessary. Without this information, national performance objectives may be met even as delays in contact investigation steps result in on-going transmission or complications of delayed TB diagnosis. An aggregate contact investigation report that could capture all the data necessary to inform potential interventions would be unwieldy and virtually impossible to implement. An alternative, therefore, is to develop or adapt existing computer information systems to allow jurisdictions to capture and analyze individual patient data and to conduct real-time evaluation of contact investigations. Several of the highest morbidity jurisdictions have computer information systems to aid in managing contact investigations, evaluate these activities, and generate the ARPE. However, at present, data elements and definitions are not standardized across these jurisdictions. Most California jurisdictions use a paper contact management system and expend considerable resources to manually review charts to complete the ARPE. These jurisdictions have requested state assistance in developing an information system to facilitate contact management and evaluation, and to generate the ARPE. Development and implementation of a web-based system that integrates case and contact information may provide an opportunity to standardize data collected for contact investigations and to encourage collection of essential information to inform interventions. To address the critical information gaps discussed above, TBCB has recently begun implementation of a CDC-funded contact investigation improvement project, in collaboration with one jurisdiction. This project will include development, implementation, and evaluation of a model data collection form and database. These tools are intended to enhance evaluation of contact investigation activities, address information and information system gaps, and facilitate design of evidencebased interventions. Tools and interventions may be replicated in other jurisdictions, as appropriate. Enhancing information collected and providing jurisdictions with tools to manage contact investigation information will assist in evaluating contact investigations and designing evidence-based interventions. TBCB has also recently initiated an intervention to improve contact investigation performance in higher morbidity jurisdictions. As part of an overall evaluation of jurisdictions TB prevention and control activities, four contact investigation measures (contact elicitation, evaluation, treatment start, and treatment completion) were developed as indicators of contact investigation performance and statewide objectives were set (Table 2). These indicators may be useful in identifying performance gaps at the state and local levels. However, if suboptimal performance is identified, the ability of jurisdictions without contact information systems to design evidence-based interventions is impaired; similarly, TBCB s ability to develop state-level interventions is limited. In conclusion, improving contact investigation practices through evidence-based interventions in a state that accounts for approximately one fifth of the nation s contact investigations may have a substantial impact on national performance. Many resources within California are dedicated to contact investigations, but our efforts are impaired because we do not know whether contacts at greatest risk of LTBI and TB disease are elicited, evaluated, and prioritized to complete treatment for LTBI; nor do we know the reasons why contacts are lost during the contact investigation process. Without this information and a statewide information system to collect standardized data, our ability to formulate and evaluate interventions is substantially undermined. Unless we address these fundamental deficiencies, we will be unable to maximize contact investigation performance and make progress in breaking the cycle of M. tuberculosis transmission. Acknowledgements We thank L Johnson for assistance in creating the ARPE dataset and the many individuals in California s local TB control programs who collected, aggregated, and reported ARPE data. References 1 Centers for Disease Control and Prevention. Reported Tuberculosis in the United States, Atlanta, GA: US Department of Health and Human Services, CDC, July 1995 September California Department of Health Services. Report on Tuberculosis in California, Berkeley, CA: California Department of Health Services. June California Department of Health Services/California Tuberculosis Controllers Association. Joint Guidelines for Contact Investigations in California. Berkeley, CA: California Department of

6 S368 The International Journal of Tuberculosis and Lung Disease Health Services/California Tuberculosis Controllers Association, 11/12/98. 4 Centers for Disease Control and Prevention. Aggregate Reports for Tuberculosis Program Evaluation: Follow-Up and Treatment for Contacts to Tuberculosis Cases. OMB No Atlanta, GA: CDC, Ferebee S H. Controlled chemoprophylaxis trials in tuberculosis: a general review. Adv Tuberc Res 1970; 17: Snyder D C, Chin D P. Cost-effectiveness analysis of directly observed therapy for patients with tuberculosis at low risk for treatment default. Am J Respir Crit Care Med 1999; 160: Binkin N J, Vernon A A, Simone P M, et al. Tuberculosis prevention and control activities in the United States: an overview of the organization of tuberculosis services. Int J Tuberc Lung Dis 1999; 8: Moodie A S, Riley R L. Infectivity of patients with pulmonary tuberculosis in inner city homes. Am Rev Respir Dis 1974; 110: