Completion Rates, Adverse Effects, and Costs of a 3-Month and 9-Month Treatment Regimen for Latent Tuberculosis Infection in California Inmates, 2011-2014

Abstract

Objectives:

In California, about 80% of tuberculosis disease is caused by untreated latent tuberculosis infection (LTBI), and the rate of LTBI is higher among incarcerated persons (16%) than among nonincarcerated persons (6%). We compared 2 regimens to treat LTBI in an adult prison population in California: 9 months of twice-weekly isoniazid (9H; previous standard of care) and 12 once-weekly doses of isoniazid and rifapentine (3HP; introduced in 2011).

Methods:

We evaluated the rates of completion and discontinuation caused by hepatotoxicity among randomly selected patients with LTBI prescribed the 9H regimen in 2011 and among patients with LTBI prescribed the 3HP regimen who entered California prisons during September 2013–March 2014. We compared the cost per fully treated patient for the 2 regimens.

Results:

Of 92 patients treated with the 9H regimen, the treatment completion rate was 42% and discontinuation due to hepatotoxicity was 14%. Of 122 patients who accepted the 3HP regimen, the completion rate was 90% and discontinuation due to hepatotoxicity was 2%. The cost per fully treated patient for the 9H regimen was $981 and for 3HP was $652.

Conclusions:

In an incarcerated population, the 3HP regimen had a higher completion rate, lower hepatotoxicity, and lower cost per fully treated patient than the 9H regimen. If coupled with a high treatment initiation rate, the high rate of LTBI treatment completion with 3HP may contribute to reducing tuberculosis morbidity in California.

In December 2011, the Centers for Disease Control and Prevention (CDC) endorsed a new short-course treatment regimen for latent tuberculosis infection (LTBI), 3HP, which consists of 12 once-weekly doses of rifapentine and isoniazid.¹ The previous standard of care for LTBI treatment was 9 months of isoniazid (9H). Although the 3HP regimen is shorter in length (and, thus, more likely to be completed) than the 9H regimen, its higher cost is a barrier to implementation.

In 2017, the LTBI rate among inmates under the care of California Correctional Health Care Services (CCHCS)—the agency that provides health care to adult inmates in state correctional institutions (ie, prisons) in California—was 16%, which was higher than the most recent estimate of 6% for the general California population in 2011.² Because of the high rate of LTBI among inmates and because untreated LTBI results in 80% of tuberculosis disease in California,³ LTBI treatment is a priority for CCHCS. In 2012-2014, we assessed the impact on treatment completion, treatment toxicity, and cost of using 3HP instead of 9H to treat LTBI for CCHCS patients.

We performed 3 separate evaluations among CCHCS patients (adult inmates residing in correctional institutions in California). In our first evaluation, we assessed a cohort of patients treated with 9H to (1) determine the completion rate of the 9H regimen and reasons for discontinuation and (2) estimate the likely completion rate for 3HP based on the proportion of patients who completed 12 weeks of 9H. In our second evaluation, we compared the cost per fully treated patient for the 9H and 3HP regimens. In our third evaluation, we assessed a group of patients who were treated with the 3HP regimen and determined the actual completion rate, all adverse events, and reasons for discontinuation of treatment. To our knowledge, this study is the first to assess the use of the 3HP regimen in a prison population and the first to compare the 3HP and 9H regimens by cost per fully treated patient in an institutional setting.

Methods

9H Retrospective Study

To assess the completion rates and reasons for discontinuation of 9H, we performed a retrospective study in March 2012. CCHCS maintains a registry of patients being treated for LTBI, derived from pharmacy data that record the number of weeks that LTBI medication has been dispensed (unpublished data, Quality Assurance Branch, CCHCS, March 2012). We defined 38 weeks of 9H treatment as completed. For the study cohort, we randomly selected 100 of 338 patients whose last isoniazid dispense date was in October or November 2011. We used this cutoff to ensure that all patients either completed the regimen or had time to complete the regimen (because the 9 months of treatment can be provided during a 1-year time frame for patients who miss doses). We performed chart reviews for all cohort patients to determine the number of weeks of treatment, the proportion of patients who completed treatment (ie, had at least 38 weeks of treatment), and the reasons for discontinuation among patients who did not complete treatment. We excluded patients whose chart and registry records did not match.

Cost Analysis

Before being prescribed treatment, patients in whom LTBI has been detected (through a positive tuberculin skin test or a positive interferon gamma release assay [IGRA]) undergo a chest x-ray, baseline liver function tests, and assessment by a registered nurse and a physician. In CCHCS, LTBI treatment requires direct observation by a licensed vocational nurse, during which the nurse watches the patient swallow the medication. For the 9H regimen, a registered nurse evaluates patients for adverse drug effects once per month. For the 3HP regimen, we projected that a registered nurse would assess for adverse effects once per week and that clinicians would order liver function testing of patients once per month. Patients on either regimen undergo monthly liver function tests.

We used 2012 pricing for our estimates. Medication costs were the average wholesale prices of isoniazid and rifapentine (CCHCS does not receive federal assistance for LTBI treatment costs), which were $1.02 per week for the 9H regimen and $21.27 per week for the 3HP regimen. The cost for a liver function test was $20 and for a chest x-ray was $50. We derived the hourly wages for physicians and nurses from state salary tables and added 30% of their salary for the cost of benefits. The hourly wage estimates were $157.30 for physicians, $45.50 for registered nurses, and $32.50 for licensed vocational nurses. We estimated the initial visit to include 15 minutes of physician time, 20 minutes of registered nurse time, and 5 minutes of licensed vocational nurse time. We estimated the time required for directly observed therapy administration as 2 minutes of licensed vocational nurse time (twice per week for the 9H regimen and once per week for the 3HP regimen) and for symptom screening as 15 minutes of registered nurse time (once per month for the 9H regimen and once per week for the 3HP regimen).

We calculated the cost of a single completed regimen for 9H and 3HP from the itemized medication, laboratory, chest x-ray, and clinician costs (Table 1). We also estimated the cost per fully treated patient for each regimen by using data from the 9H retrospective cohort. We defined this cost as the cost of treating all patients divided by the number of fully treated patients. We estimated the cost of drugs and nurse time (numerator data) as the cost to treat all patients for LTBI, including patients who did not complete the regimen.

Table 1.

Comparison of estimated cost to complete a course of latent tuberculosis infection (LTBI) treatment: 38 weeks of twice-weekly isoniazid (9H) vs 12 weeks of once-weekly rifapentine and isoniazid (3HP) for 1 patient, California Correctional Health Care Services, March 2012

Itemized Required Resource	Cost of 38-Week 9H Regimen, $a	Cost of 12-Week 3HP Regimen, $a
Total medication for complete LTBI treatment	39	255
Weekly medication cost	1	21
Initial visit total	127	127
Chest x-ray	50	50
Liver function tests	20	20
LVN time (for blood draw)	3	3
RN time (for education and symptom screening)	15	15
MD time (for decision on LTBI regimen)	39	39
Total weekly clinical staff time	82	150
Clinician time per week	2	12
LVN time (for directly observed therapy)	2	1
RN time (for symptom screening)	NA	11
Total monthly visits	307	68
Per monthly visit	34	23
Liver function tests	20	20
LVN time (for blood draw)	3	3
RN time (for symptom screening)	11	NA
Total resources for a completed LTBI treatment regimen for 1 patient	555	600

Abbreviations: LVN, licensed vocational nurse; MD, medical doctor; NA, not applicable; RN, registered nurse.

^a Cost data are 2012 costs provided by the California Correctional Health Care Services pharmacy department.

We estimated the cost for the 9H regimen (numerator data) as the cost to treat all patients in the 9H cohort (the cost to treat patients who did and did not complete the 38-week treatment). We used data from the 9H cohort study to estimate the costs for the 3HP regimen. We estimated the cost for the 3HP regimen as the cost to treat all patients through 12 weeks (the cost to treat patients who did and did not complete the 12-week treatment). For example, we included in the numerator for both regimens the cost to treat a patient who completed only 10 weeks of treatment (which included initial visit costs plus weekly costs [Table 1] multiplied by the number of weeks of treatment completed), but we did not include that patient in the denominator of fully treated patients. For the 9H cohort, we defined “fully treated” (denominator) as the number of patients in the 9H retrospective cohort who completed at least 38 weeks of treatment. For the 3HP cohort, we defined “fully treated” as the number of patients in the 9H retrospective cohort who completed 12 weeks of treatment. At this stage, we had not conducted the 3HP demonstration project, so we used the 12-week completion rate among patients treated with the 9H regimen to estimate the proportion of patients who would be expected to complete 12 weeks of treatment using the 3HP regimen. We hypothesized that it would be easier for patients to adhere to the 3HP regimen than to the 9H regimen because the 3HP regimen requires only 1 dose per week (and we would expect to see for it a higher completion rate than the rate observed for 12 weeks of the 9H regimen). However, we used the completion of 12 weeks with the 9H regimen as a conservative estimate. We conducted a cost assessment before we offered the 3HP regimen to our patients; we used data on weeks of treatment completed for the 9H regimen to infer the treatment completion rate expected if we were to offer the 3HP regimen.

Because treating LTBI with 6 months of isoniazid (6 H regimen) has some benefits,⁴ we also conducted a sensitivity analysis to assess use of the 6-month regimen. Because the 6 H regimen is about 80% as effective as the 9H or 3HP regimens, we adjusted the cost per fully treated patient by dividing the cost per treated patient by 0.8.

3HP Prospective Study

In 2013, we conducted a prospective study to determine the performance of the 3HP regimen in our cohort. We defined treatment completion according to the measure used in the clinical trial¹ that contributed to CDC’s endorsement of 3HP—we considered a patient to have completed the regimen if he or she ingested 11 doses of the combination treatment within 16 consecutive weeks. Although 12 weeks is the standard and optimal regimen, the California Tuberculosis Controllers Association considers 11 weeks of 3HP treatment within 16 consecutive weeks to be complete.⁵ For the prospective study, our goal was to enroll a convenience sample of approximately 100 patients entering our system with LTBI. In 2013, inmates entered the California state prisons through 6 reception centers. Four reception centers participated in the project: the only reception center for women (reception center 1) received 2766 women during 2013; the 2 largest reception centers for men and women (reception center 2 and reception center 3) received 11 669 men and 14 357 men during 2013, respectively; and 1 of 3 small reception centers for men (reception center 4) received 2885 men in 2013 (unpublished CCHCS data, 2013).

We asked the executive teams at the 4 reception centers to complete selection forms for each patient who entered the system with LTBI (diagnosed as described under Cost Analysis) and to offer the 3HP regimen to all patients who had no record of previous treatment for LTBI by any regimen and had no evidence of tuberculosis disease. Our study was not intended to compare one group treated with 3HP with another group treated with 9H during the same time frame. After conducting the cost analysis, which indicated that the 3HP regimen would be fiscally sound, we offered the 3HP regimen to a limited number of patients who we could follow closely during their treatment (eg, with weekly evaluations). We excluded patients who were pregnant; were infected with HIV and being treated with antiretrovirals; wore dentures or contact lenses (because these items could be permanently stained by rifapentine); were taking antiepileptics or warfarin (because of the potential for drug interactions); were suspected of having active tuberculosis disease; were suspected of having LTBI resistant to isoniazid or rifapentine; or had a history of hypersensitivity to a rifamycin. After evaluating a patient, the physician at the reception center made the final determination about whether or not to enroll the patient. On offer of 3HP, nurses informed patients that treatment acceptance was voluntary.

Nurses administered the medication to patients by using directly observed therapy and evaluated each patient once per week for symptoms of adverse effects. When a patient completed or discontinued the 3HP regimen, the project registered nurse returned that patient’s weekly monitoring form to CCHCS headquarters. In addition to detailing the patient’s symptoms while on treatment, nurses documented regimen completion or the reason(s) for premature discontinuation on the monitoring forms. At the end of the project, we computed rates of completion and adverse events that led to discontinuation of treatment.

Adverse Effects

Based on chart reviews for the retrospective study and monitoring forms for the prospective study, we determined adverse effect(s) that led to discontinuation of treatment. We based our assessment of hepatotoxicity on a physician’s report of the presence or absence of symptoms and laboratory reports of the transaminase levels at the time of treatment termination. We also used the American Thoracic Society criteria for discontinuing a regimen for hepatotoxicity,⁶ specifically, aminotransferase levels more than 5 times a laboratory’s upper limit of normal for patients with no symptoms and more than 3 times the upper limit of normal for patients with symptoms. We defined hypersensitivity as having any of the following symptoms: hypotension (low blood pressure), urticaria (hives), angioedema (swelling under the skin), acute bronchospasm (sudden constriction of the muscles in the bronchioles), conjunctivitis (pink eye), or fever.

Results

9H Retrospective Study

Of 100 randomly selected patients who had initiated treatment, we excluded 8 for discrepancies between the registry and chart accounts of their LTBI regimens (Figure 1). Of the remaining 92 patients, as of March 2012, 39 (42%) had completed at least 38 weeks of the 9H regimen and 53 (58%) had discontinued the 9H regimen before completion. Of the 53 patients who discontinued treatment before completion, 32 (60%) were paroled or transferred out of the system before 38 weeks, 4 (8%) refused to complete treatment, and 17 (32%) had their regimens discontinued by providers. Of the 17 patients whose regimens were discontinued by providers, the treatment of 13 (14% of the 92 who started the 9H regimen) was discontinued because of concerns for hepatotoxicity; 7 of these met the American Thoracic Society criteria for hepatotoxicity (4 had symptoms and 3 had no symptoms).

Figure 1.

Treatment completion outcomes of 100 randomly selected patients with latent tuberculosis infection (LTBI) treated with a regimen of 9 months of isoniazid twice per week (9H), with a last dispense of isoniazid in October or November 2011, California Correctional Health Care Services, March 2012. Of regimens discontinued for reasons other than hepatotoxicity or hypersensitivity, the treatment of 1 patient was discontinued because of blurry vision and the treatment of 2 patients was discontinued because of negative results on interferon gamma release assays performed after the regimens were started.

Of regimens discontinued for reasons other than hepatotoxicity, the treatment of 1 patient was discontinued for hypotension, 1 for blurry vision, and 2 because of negative results on IGRAs performed after the regimens were started. For these 2 patients, chart reviews indicated that clinicians regarded the negative IGRAs as indicating that the patients did not need LTBI treatment.

Cost Analysis

We estimated the total cost for a single completed 9H regimen to be $555 and for a single completed 3HP regimen to be $600 (Table 1). To compute the cost per fully treated patient, we found 39 (42%) patients had completed at least 38 weeks of 9H treatment, and 79 (86%) patients had completed at least 12 weeks of 9H treatment; we used the latter as an estimate for the number of patients completing 12 weeks of the 3HP regimen (Figure 2). The total cost of treating the full cohort of 92 patients with 9H was $38 243 ($16 589 for 53 patients treated <38 weeks and $21 654 for 39 patients treated ≥38 weeks) (Table 2). The total cost of treating the full cohort of 92 patients with the 3HP regimen was $51 533 ($4125 for 13 patients treated for <12 weeks and $47 408 for 79 patients who completed treatment). In our review of 92 patients treated with the 9H regimen in 2011, 79 (86%) completed ≥12 weeks of treatment; as such, we estimated that 86% of patients would complete the 3HP regimen.

Figure 2.

Proportion of patients with latent tuberculosis infection (LTBI) treated with a regimen of 9 months of isoniazid twice per week (9H) whose last isoniazid dispense date was in October or November 2011 (N = 92), by number of weeks of treatment completed, California Correctional Health Care Services, March 2012. Of 92 patients who started on the 9H regimen, 86% completed ≥12 weeks of treatment; had they been on the 3HP regimen (12 weeks of isoniazid and rifapentine), 12 weeks would be the time needed to complete treatment. Thus, 79 (86%) patients would be expected to be fully treated with the 3HP regimen by week 12 (black bar at 12 weeks). Of the 92 patients who started on the 9H regimen, 39 (42%) completed the full course of treatment (ie, ≥38 weeks) (black bar at 38 weeks).

Table 2.

Estimated cost per fully treated patient^a with latent tuberculosis infection for 38 weeks of twice-weekly isoniazid (9H regimen) compared with 12 weeks of once-weekly rifapentine and isoniazid (3HP regimen), California Correctional Health Care Services, March 2012^b

Measure	9H Regimen	3HP Regimen
No. (%) of patients who completed LTBI treatment (among those who initiated treatment, N = 92)	39 (42)	79 (86)
Cost for patients who were partially treated (<38 weeks for 9H group, n = 53; <12 weeks for 3HP group, n = 13), $c	16 589	4125
Cost for patients who completed LTBI treatment (n = 39 for 9H group; n = 79 for 3HP group), $c	21 654	47 408
Cost to treat the full cohort (cost to treat patients who were partially treated plus cost to treat patients who completed treatment) (N = 92 patients), $c	38 243	51 533
Cost per fully treated patient (cost to treat the full cohort [N = 92] divided by the number of patients who completed treatment), $c	981	652

^a Cost per fully treated patient is the cost to treat patients who did and did not complete the regimen (eg, completed only 10 weeks of treatment) divided by the number of patients who completed the regimen. If all patients completed the regimen, the cost per fully treated patient would be the same as the cost of the regimen for 1 patient ($555 for 9H and $600 for 3HP).

^b Data for this analysis were based on data from 92 patients who completed the 9H regimen in 2011. The estimated proportion of patients who completed the 9H regimen is the number of the 92 patients who completed ≥38 weeks of the 9H regimen. The estimated proportion of patients who completed the 3HP regimen (12 weeks of treatment) is the number of the 92 patients who completed ≥12 weeks of the 9H regimen.

^c Cumulative costs for partially or fully treated patients were obtained by summing individual treatment costs for each patient who partially or fully completed the indicated treatment regimen. Individual treatment costs were calculated by summing the costs for the initial visit (Table 1) and weekly treatment costs multiplied by the number of weeks of treatment completed by each patient.

The cost per fully treated patient for the 9H regimen was $38 243 divided by 39 patients who completed ≥38 weeks of treatment, or $981. For the 3HP regimen, the total cost per fully treated patient was $51 533 divided by 79 patients who completed ≥12 weeks of treatment, or $652.

In our sensitivity analysis for completion of the 6 H regimen, the cost to fully treat 1 patient was $616. The number of fully treated patients was 55, and the number of partially treated patients was 45. When adjusting for the lower efficacy of the 6 H regimen compared with the 9H or 3HP regimen, the cost per fully treated patient was $780.

3HP Prospective Study

From September 3, 2013, through March 28, 2014, 352 patients with LTBI were considered for the 3HP demonstration project. All 352 patients were offered treatment, and 122 (35%) accepted treatment. Accepted patient counts and proportions among patients considered for enrollment were 0 of 36 at reception center 1, 75 of 242 (31%) at reception center 2, 33 of 43 (77%) at reception center 3, and 14 of 33 (42%) at reception center 4.

One hundred ten of 122 (90%) patients enrolled in the 3HP regimen completed treatment. Completion rates ranged from 85% to 100% at the reception centers. An analysis of weekly monitoring forms found that 30 of 122 (25%) patients who were selected for the project reported at least 1 adverse symptom during treatment. The most common symptom, headache, was reported by 13 (11%) patients. No patients were hospitalized and none died.

Of the 12 patients who discontinued treatment before completion, 5 (42%) left the prison system before completing the regimen, 3 (25%) refused to continue treatment, and 4 (33%) had their regimen discontinued by a provider (Figure 3). Providers discontinued the regimen for 2 patients because of hepatotoxicity (2% of 122 enrolled patients), but only 1 of these patients had symptoms and liver function test results that met the American Thoracic Society criteria for hepatotoxicity. Of the 2 additional patients whose regimens were discontinued by a provider, 1 had no symptoms after the first 4 doses of the regimen but developed severe nausea and vomiting after the fifth dose, and 1 had no symptoms after the first 2 doses of the regimen but had subjective fever and chills after the third dose.

Figure 3.

Treatment completion outcomes of patients who accepted treatment for latent tuberculosis infection (LTBI) with the 3HP (12 doses of once-weekly isoniazid and rifapentine) regimen, California Correctional Health Care Services, September 2013–March 2014 (n = 122).

Discussion

Our evaluation found that completion with the 3HP regimen was more than twice as high as with the 9H regimen (90% vs 42%). The high completion rate for the 3HP regimen is comparable with or higher than 3HP completion rates reported in other published studies. In a multicenter trial of persons at high risk for tuberculosis disease, the completion rate for the 3HP regimen was 82%⁷; in a county jail population, the completion rate was 85%⁸; in a community health center serving a low-income Hispanic population, the completion rate was 78%⁹; among persons diagnosed with LTBI at 2 New York City Department of Health and Mental Hygiene Tuberculosis Clinics, the completion rate was 65%¹⁰; and in tuberculosis treatment programs across the United States, the completion rate was 87%.¹¹

We also found the 3HP regimen to be more cost effective than the 9H regimen when we evaluated the cost per fully treated patient. Although these estimates were for patients in correctional institutions, they would likely be similar to the per-patient costs in other institutionalized populations (eg, mental health institutions). Despite medication costs that were more than 6 times higher for the 3HP regimen than for the 9H regimen, the 3HP regimen was only 8% more expensive than the 9H regimen when we included health care staff member time in the calculations. Our cost estimates were conservative, and the 3HP regimen would have shown more cost savings had we added in other expenses, such as the costs of correctional officers escorting our patients to clinician appointments and the costs of responding to hepatotoxicity. Unlike in the community, where the directly observed therapy requirement for the 3HP regimen is inconvenient and expensive compared with self-administered isoniazid,¹⁰ administration of directly observed therapy is common and easily implemented in institutional settings where medications are often administered directly by a nurse. In CCHCS, all LTBI treatment regimens are administered by using directly observed therapy, and all regimens include frequent visits with health care staff members, making a shorter duration of 3HP treatment more convenient and cost effective than the 9H regimen. In addition, tracking patients (who frequently move between institutions) is less prone to error when using a short-course regimen, because a change in the case manager is less likely than with a longer regimen.

Discontinuation for hepatotoxicity was more common among patients on the 9H regimen (14%) than among patients on the 3HP regimen (2%), a finding that others have reported in comparisons of patients on the 3HP and 9H regimens.^6,12 Even after we applied the American Thoracic Society criteria for discontinuing a regimen for hepatotoxicity,¹³ patients on the 9H regimen were more likely than patients on the 3HP regimen to have treatment discontinued because of hepatotoxicity (9H: 7 of 92 [8%]; 3HP: 1 of 122 [1%]). Using these stricter criteria ensures that the comparison is more objective and less subject to differences in the assessment of different providers, given that the 2 assessments were conducted in different years and the medical providers likely differed.

CCHCS needs to train clinicians to recognize that carefully monitored inmates with rising aminotransferase levels can complete LTBI treatment and that the presence of hepatitis C virus antibody is not necessarily associated with LTBI treatment hepatotoxicity.^14,15 Physicians should suspend treatment for hepatotoxicity in accordance with American Thoracic Society guidelines and then attempt to resume treatment with the same or another medication once symptoms resolve and liver function tests normalize.

Some providers also discontinued the regimen among patients who tested positive with a tuberculin skin test and subsequently tested negative with an IGRA test. Although a negative IGRA result does not indicate definitively that a patient does not have LTBI and false-negative IGRAs occur, no definitive confirmatory test is available; if results differ between the IGRA and the tuberculin skin test, the decision to continue treatment is left to clinical judgment.

We are encouraged by the high completion rate with the 3HP regimen. The access to care for CCHCS patients, which is likely better than access to care in the community, provides an opportunity to treat LTBI infections and to prevent tuberculosis disease in the community. In California, only about 20% of persons with LTBI know that they are infected and only about 12% have been treated.^2,3 In contrast, because LTBI screening is required in the California prison system, all CCHCS patients with LTBI know that they are infected, and all patients with untreated LTBI are encouraged to start treatment every year during the required annual tuberculosis screening that occurs during their birth month. Because the average incarceration length in California is 36 months and almost all patients are in prison for at least 6 months, most patients are able to complete the 3HP regimen during their incarceration.¹⁶

Limitations

This study had several limitations. First, for the retrospective study, we measured treatment completion by weeks of treatment rather than by doses of treatment, the accepted standard of assessing adherence. However, we believe that weeks of treatment was a good approximation of completion; this hypothesis was borne out by the similarity between our estimate of the proportion of patients who would likely complete 12 weeks of the 3HP regimen (based on the proportion of patients who completed 12 weeks of the 9H regimen [86%]) and the proportion of patients who actually completed the 3HP regimen in the prospective study (based on at least 11 doses during a 16-week period administered by using directly observed therapy [90%]). Second, we did not allow for the possibility of continuation and completion of treatment outside the prison system (ie, for patients who were released or paroled from prison). CCHCS has no jurisdiction over persons released from the prison system, and treatment adherence in the community cannot be tracked. However, a review of the literature indicates that inmates are unlikely to continue treatment after discharge, even for tuberculosis disease.^17,18 Third, our cost analysis did not include the cost savings or quality-adjusted life years gained by preventing tuberculosis disease, nor did it give credit for the 6 H regimen, which has been shown to be somewhat effective.¹⁸ Nevertheless, only 60% of patients in our study completed even 6 months of the 9H regimen, and because the 3HP regimen was similar in cost to, less hepatotoxic than, and noninferior to the 9H regimen,¹ there would be no reason to prefer the 9H regimen in this population. Fourth, we were unable to enroll any women; the single reception center for women reported that all women entering with LTBI had already been treated. Thus, the study was not generalizable to women in our system. Lastly, the sample size was relatively small and the 9H study was somewhat dated, although we would not expect a difference in adherence to the 9H regimen used today versus used in 2011 (because the main reasons for discontinuation, such as patient fatigue with the long regimen, being paroled, and toxicity, did not change in this interval).

Conclusions

Use of the 3HP regimen promoted a high rate of LTBI treatment completion among incarcerated persons in California. Because the 3HP regimen is less hepatotoxic than the 9H regimen, this regimen may be particularly beneficial for populations at high risk for liver disease (eg, because of hepatitis C virus infection). The short-course regimen permitted completion of treatment before parole or discharge. Unlike with the longer regimen, patients did not need to rely on a health care system in the community, which can be difficult to access, to complete treatment. Our findings of safety and high completion rates may encourage use of the 3HP regimen by other correctional health care systems.

Because patients who have been incarcerated may not receive adequate access to care in the community after incarceration, health care systems that ensure access to health care in prisons can positively affect both the individual health of patients and the health of the community. Persons who complete LTBI treatment while incarcerated are much less likely to develop tuberculosis disease and transmit tuberculosis disease to the community than persons who do not complete LTBI treatment. In addition, because 80% of tuberculosis disease stems from untreated LTBI in California, and because LTBI rates are higher in the incarcerated population than in the community in California, the high LTBI treatment completion rate with the 3HP regimen, coupled with high acceptance rates of the regimen, should contribute to reducing the burden of tuberculosis disease in California.