Guide to the Application of Genotyping to Tuberculosis Prevention and Control

Applying Genotyping Results to Tuberculosis Control Practices

Cluster Investigations

Goal

The goal of a cluster investigation is to uncover previously unrecognized epidemiologic links between persons in a genotyping cluster and to take appropriate action steps.

Steps

Cluster investigations involve taking multiple steps and collecting information from multiple sources. Detailed cluster investigation protocols and data collection forms have been developed by several state TB programs and are available for review and downloading at the CDC WebBoard at http://web-tb.forum.cdc.gov under TB Genotyping. Table 6.2 lists some of the important steps to consider when conducting a cluster investigation. For each step, it will be important to assign responsibility to either the state or the local TB program.

After the decision to launch a cluster investigation is communicated to the local programs that managed the patients in the cluster, the next key step is to analyze existing information about the patients in the cluster in order to determine what is already known about how the patients might be related. In most cases, the local program should take the lead on reviewing existing information about all persons in a cluster. This information can come from a review of existing information already collected about the persons (e.g., the initial case interviews and contact investigations) and from talking with the case managers, epidemiologists, or outreach workers who know the patients.

The state and local programs should work together to analyze the results of the existing information to determine if interviews with the patients themselves are needed. If this initial data abstraction and interview process identifies a) epidemiologic links between all cases in the cluster, b) the direction of transmission from source to secondary cases, and c) where the transmission probably occurred, persons in the cluster may not need to be interviewed. If the cluster is a high priority to investigate and it is decided that a patient interview is required, it is usually best for the local program to take the lead on conducting patient interviews. On the other hand, if local staffing is insufficient, the state program staff may need to help out.

Table 6.2. Suggested roles for state and local TB program staff in cluster investigations.

If a decision is made to interview the patients in a cluster, the existing information gathered and analyzed during the previous steps will be helpful in guiding the subsequent interviews. Interview forms are also helpful, and a template of a cluster investigation form (TB Cluster Investigation Form.doc) is available from the CDC WebBoard.

Although any information that is gained may hold the key to understanding how transmission between persons in the cluster might have occurred, the most valuable information to collect often involves the persons’ social history, their behaviors, where they spend their leisure time, church and school attendance, what health-care facilities they have visited, foreign travel, work history, incarceration, whether they are residents of a nursing home or a single-resident occupancy hotel, or if they have slept at homeless shelters. The time period of focus for the cluster investigation is usually longer than for a contact investigation. In order to understand the possible transmission dynamics of cases that led to the present cluster, questions about behaviors and leisure-time activities should seek information for the preceding 5 years, even though recent transmission is defined as occurring within the previous 2 years.

After the first patient is interviewed in the cluster investigation, the person conducting the interviews should update the interviewer notes to take into account any new information gained during the first interview. The locations, activities, or behaviors identified during the first interview should become prompts for subsequent interviews. Questions about specific locations, however, should not be asked during subsequent interviews in order to ensure confidentiality.

Outcome

If known epidemiologic links have not been identified between the persons with matching genotypes, no further action is required at this point. Of course, saying that no epidemiologic links were identified is not the same as concluding that no epidemiologic links exist, and the absence of links does not mean that recent transmission did not occur. If TB patients who have a matching genotype that adds them to this cluster are identified in the future, the chances increase that an undiscovered epidemiologic link exists. In this case, a careful analysis of possible links that might explain all the patients in a cluster should be undertaken.

If the cluster investigation has been able to establish known epidemiologic links between persons in the cluster, the TB program should decide if these linked persons represent an outbreak and if an outbreak investigation is indicated. Alternatively, if an outbreak is not thought to be occurring, the TB program probably will still want to expand the original contact investigation to include contacts that might share the same link that exists between the clustered patients.