Date of Award

January 2018

Document Type


Degree Name

Medical Doctor (MD)



First Advisor

Frederick L. Altice


Our two simultaneous studies in a Peruvian prison aimed to (1) find factors associated with 100% adherence to tuberculosis (TB) treatment, and (2) use one-on-one in-depth interviews to explore whether and why a unique peer-led prison TB program was successful from the perspective of different stockholders. These two studies were mixed in the data analysis face to better inform each independent set of results.

A surveying and medical record review of 251 prisoner patients who underwent TB treatment within the last year was done to assess participant demographics, HIV status, alcohol dependence, depression, stigma towards TB, adherence to TB treatment, and impression of care providers. Bivariate and multivariate analyses were done on two independent variables: 100% adherence and loss of trust in health care staff. Also, nine interviews with healthcare staff were conducted, coded, and analyzed using a method of constant comparison to formulate key themes and an overarching narrative. Our study was then mixed in the data analysis phase to (1) complete a community narrative and include prisoners who were not interviewed, and (2) to contextualize our survey results.

Our results showed a rate of 100% adherence, defined as not missing a single day of Directly Observed Treatment (DOT) through patient self-report, was 77.7%. Our multivariate model for 100% adherence showed that having a guard know your TB status was positively associated with DOT 100% adherence (aOR: 2.77, CI: 1.33 – 5.75), while having either a doctor (aOR: 0.45, CI: 0.23 – 0.88) or prisoner (aOR: 0.38, CI: 0.20 – 0.73) as your DOT Observer was associated with DOTS non-100% adherence. Our model for loss of trust showed that fear toward TB treatment (aOR: 2.86, CI: 1.05 – 7.91) and reported mistreatment by medics due to TB status (aOR: 16.47, CI: 5.44 – 49.84) were positively associated with loss of trust.

Our interviews describe a setting with insufficient resources that rely on prisoner labor to offset a shortage of staff. A peer-based intervention (PBI) trains prisoners, named Promotores to screen fellow prisoners for TB, provide education, and assist in treatment delivery. The description of challenges faced in implementing this program gives a narrative of dual competing interest: Physician and nurses aim to provide medical care for their patients, while promoting order and security in a setting where this is not guaranteed. Promotores similarly split their focus, balancing work for the TB program and survival in a prison that does not guarantee adequate food, shelter or security. In seeking more resources, Promotores, at times, abuse their privileges and their fellow prisoners. Advice and future directions for implementing a peer-based intervention of this type are given.

The PBI program described in this study is unlike any in the peer-reviewed or grey literature in its focus on TB. Its implementation has mixed results: prisoner labor allows the existence of key screening techniques, and the results are widely praised by health workers; yet, abuses of power are discovered and the voices of prisoner patients are absent from our analysis. Further research is needed in this setting.


This thesis is restricted to Yale network users only. It will be made publicly available on 06/25/2100