Jessie K Edwards1, Yeycy Donastorg2, Sabrina Zadrozny3, Sarah Hileman4, Hoisex Gómez2, Marissa J Seamans5, Michael E Herce6, Edwin Ramírez7, Clare Barrington8, Sharon Weir1,8. 1. From the Department of Epidemiology, University of North Carolina at Chapel Hill, Chapel Hill, NC. 2. Instituto Dermatológico y Cirugia de Piel, Santo Domingo, Dominican Republic. 3. Frank Porter Graham Institute, University of North Carolina at Chapel Hill, Chapel Hill, NC. 4. Independent, Chapel Hill, NC. 5. Department of Epidemiology, University of Los Angeles, Los Angeles, CA. 6. School of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, NC. 7. Servicio Nacional de Salud, Santo Domingo, Dominican Republic. 8. Carolina Population Center, University of North Carolina at Chapel Hill, Chapel Hill, NC.
Abstract
BACKGROUND: Improving viral suppression among people with HIV reduces morbidity, mortality, and transmission. Accordingly, monitoring the proportion of patients with a suppressed viral load is important to optimizing HIV care and treatment programs. But viral load data are often incomplete in clinical records. We illustrate a two-stage approach to estimate the proportion of treated people with HIV who have a suppressed viral load in the Dominican Republic. METHODS: Routinely collected data on viral load and patient characteristics were recorded in a national database, but 74% of patients on treatment at the time of the study did not have a recent viral load measurement. We recruited a subset of these patients for a rapid assessment that obtained additional viral load measurements. We combined results from the rapid assessment and main database using a two-stage weighting approach and compared results to estimates obtained using standard approaches to account for missing data. RESULTS: Of patients with recent routinely collected viral load data, 60% had a suppressed viral load. Results were similar after applying standard approaches to account for missing data. Using the two-stage approach, we estimated that 77% (95% confidence interval [CI] = 74, 80) of those on treatment had a suppressed viral load. CONCLUSIONS: When assessing the proportion of people on treatment with a suppressed viral load using routinely collected data, applying standard approaches to handle missing data may be inadequate. In these settings, augmenting routinely collected data with data collected through sampling-based approaches could allow more accurate and efficient monitoring of HIV treatment program effectiveness.
BACKGROUND: Improving viral suppression among people with HIV reduces morbidity, mortality, and transmission. Accordingly, monitoring the proportion of patients with a suppressed viral load is important to optimizing HIV care and treatment programs. But viral load data are often incomplete in clinical records. We illustrate a two-stage approach to estimate the proportion of treated people with HIV who have a suppressed viral load in the Dominican Republic. METHODS: Routinely collected data on viral load and patient characteristics were recorded in a national database, but 74% of patients on treatment at the time of the study did not have a recent viral load measurement. We recruited a subset of these patients for a rapid assessment that obtained additional viral load measurements. We combined results from the rapid assessment and main database using a two-stage weighting approach and compared results to estimates obtained using standard approaches to account for missing data. RESULTS: Of patients with recent routinely collected viral load data, 60% had a suppressed viral load. Results were similar after applying standard approaches to account for missing data. Using the two-stage approach, we estimated that 77% (95% confidence interval [CI] = 74, 80) of those on treatment had a suppressed viral load. CONCLUSIONS: When assessing the proportion of people on treatment with a suppressed viral load using routinely collected data, applying standard approaches to handle missing data may be inadequate. In these settings, augmenting routinely collected data with data collected through sampling-based approaches could allow more accurate and efficient monitoring of HIV treatment program effectiveness.
Authors: Nanina Anderegg; Leigh F Johnson; Elizabeth Zaniewski; Keri N Althoff; Eric Balestre; Matthew Law; Denis Nash; Bryan E Shepherd; Constantin T Yiannoutsos; Matthias Egger Journal: AIDS Date: 2017-04 Impact factor: 4.177
Authors: Nicholas A Medland; James H McMahon; Eric P F Chow; Julian H Elliott; Jennifer F Hoy; Christopher K Fairley Journal: J Int AIDS Soc Date: 2015-11-30 Impact factor: 5.396
Authors: Kristin Brown; Daniel B Williams; Steve Kinchen; Suzue Saito; Elizabeth Radin; Hetal Patel; Andrea Low; Stephen Delgado; Owen Mugurungi; Godfrey Musuka; Beth A Tippett Barr; E Amaka Nwankwo-Igomu; Leala Ruangtragool; Avi J Hakim; Thokozani Kalua; Rose Nyirenda; Gertrude Chipungu; Andrew Auld; Evelyn Kim; Danielle Payne; Nellie Wadonda-Kabondo; Christine West; Elizabeth Brennan; Beth Deutsch; Anteneh Worku; Sasi Jonnalagadda; Lloyd B Mulenga; Kumbutso Dzekedzeke; Danielle T Barradas; Haotian Cai; Sundeep Gupta; Stanley Kamocha; Margaret A Riggs; Karampreet Sachathep; Wilford Kirungi; Joshua Musinguzi; Alex Opio; Sam Biraro; Elizabeth Bancroft; Jennifer Galbraith; Herbert Kiyingi; Mansoor Farahani; Wolfgang Hladik; Edith Nyangoma; Choice Ginindza; Zandile Masangane; Fortune Mhlanga; Zandile Mnisi; Pasipamire Munyaradzi; Amos Zwane; Sean Burke; Felix B Kayigamba; Harriet Nuwagaba-Biribonwoha; Ruben Sahabo; Trong T Ao; Chiara Draghi; Caroline Ryan; Neena M Philip; Fausta Mosha; Aroldia Mulokozi; Phausta Ntigiti; Angela A Ramadhani; Geoffrey R Somi; Cecilia Makafu; Veronicah Mugisha; Julius Zelothe; Kayla Lavilla; David W Lowrance; Rennatus Mdodo; Elizabeth Gummerson; Paul Stupp; Kyaw Thin; Koen Frederix; Stefania Davia; Amee M Schwitters; Stephen D McCracken; Yen T Duong; David Hoos; Bharat Parekh; Jessica E Justman; Andrew C Voetsch Journal: MMWR Morb Mortal Wkly Rep Date: 2018-01-12 Impact factor: 17.586