Many patients who are prescribed treatment for latent tuberculosis (LTBI) do not complete treatment. Effective interventions that are low-cost and feasible for large-scale implementation are needed to support efforts to eliminate TB. The investigators hypothesize that a 3-minute educational LTBI video intervention will improve treatment completion compared to standard care (no intervention).

A randomized trial will be conducted at Kaiser Permanente Southern California, including adults who are prescribed treatment for LTBI, identified using an electronic health record algorithm. At the time of treatment prescription, patients will be randomized to receive the LTBI video intervention or standard care at a ratio of 2 to 1. Those in the intervention group will be sent a secure text message or e-mail with a link to the video. Individuals who watch the LTBI video will be invited to respond to 4 short questions about their perceptions of the video accessibility and content.

The primary outcome is LTBI treatment completion (dispensation of the required number of doses) in the year following prescription order. Secondary outcomes are LTBI treatment initiation (at least one dispensation) in the year following prescription order; initiation and completion within the year following prescription order stratified by LTBI treatment regimen (if feasible); the proportion of individuals randomized to the intervention group who watched the LTBI video; and the perceptions of those who watched the video.

The primary analysis will be based on intention-to-treat. Characteristics of individuals in the intervention and standard care groups will be described, and rates of LTBI treatment completion will be calculated. To assess the effect of the intervention, Poisson regression with robust standard errors will be used to estimate unadjusted and adjusted rate ratios and 95% confidence intervals. Per-protocol analyses will also be conducted, comparing those in the intervention group who watched the LTBI video to those in the standard care group. Similar methods will be used for secondary analyses.

Assuming the completion rate in the standard care group is 0.5 (based on recent KPSC data), approximately 918 patients will be needed to achieve 80% power with type one error 0.05 to detect an intervention effect of 0.10 in the completion rate between groups. This sample size will be feasible to achieve over the 1 year trial duration.

Please see the full protocol for details of the study.

Inclusion criteria:

• Active KSPC member

• Age ≥18 years

• Receipt of an LTBI treatment prescription order (medication ID codes listed below)

  • Isoniazid: 12199066, 1214000023, 1216000272, 12022505, 12035328, 12035329, 12035330, 12035331, 12035332, 12035333, 12035334, 12035335, 12035336, 12035337, 12015833, 12032230, 12035338, 12035339, 12035340, 12035341, 12035342, 12035343, 12035344, 12035345, 12024985, 12035346, 12181973
  • Rifampin: 12199042, 121103412, 1214000029, 1214000030, 1214000340, 1214000565, 1215200240, 1206096, 12017860, 1209060, 12011205, 12015920, 190716, 1902899, 190207268, 190210700, 12189972, 12189977, 12189987
  • Isoniazid (see codes above) + Rifampin (see codes above), also: 12032976, 12199041, 12029484, 12189973, 12168828
  • Isoniazid (see codes above) + Rifapentine (121100982, 1206442, 12025808, 12188784, 12189978)

• LTBI ICD-10 diagnosis code prior to prescription order: R76.11, R76.12, Z22.7

Exclusion Criteria:

• Active TB ICD-10 diagnosis code prior to prescription order: A15, A17, A18, A19

We did not collect information on adverse events as part of this study.

1. A small proportion of patients may have preferred other languages than English and Spanish.
2. There may have been unmeasured differences between the control and intervention groups.
3. We relied on pharmacy dispensing records and cannot confirm if all treatment doses were taken as prescribed.