Rigor and relevance in implementation research: having our cake and eating it too

In a PLOS Medicine commentary, I argue that implementation research demands relevance without sacrificing rigor (2017). With colleagues Stefan Baral, James Hargreaves and Maya Petersen, I recently edited a supplement in the Journal of AIDS that highlights contemporary applied implementation research in the HIV.

We do a lot of randomized trials in HIV – where does implementation research fit in?

Randomized trials can contribute critical insights, but in a pair of notes in PLOS Medicine, I argue we need more. Trials must account for patient preferences and emphasize external validity to be useful.

Metrics and measures in public health

Comparative effectiveness is important in some cases, but can blind us to the potential contribution of a novel interventions if we aren’t careful. A less effective intervention can still advance population health. With my Dave Glidden, we use the concept of mosaic effectiveness to explain how.

Person centered public health – from paradox to opportunity

Public health seeks to deliver services to populations, and so by definition can’t be personalized right? With Charles Holmes, Mosa Moshabela and others, I argue in a commentary as well as in a presentation at CROI, some forms of personalization are not only possible, but might be requisite for improving public health.

Studies of Health Systems Change

With partners at Makerere University in Uganda, we were among the first to apply Larry Green’s PRECEDE-PROCEED framework for health promotion to improve HIV treatment initiation using a stepped wedge cluster randomized trial.

Despite the fact that many innovations that are evidence based in medicine, some do diffuse relatively rapidly, but still require managed efforts to get to acceptable levels, was the case of immediate HIV treatment identified in a randomized trial (ACTG 5164) into real world practice at San Francisco General Hospital.

Bringing the Patient Voice into Public Health Service Delivery

We have used a range of methods to better grasp the lived experience of persons living with HIV and getting treatment and use those insights to inform services.

We developed a conceptual framework for retention (PLOS Medicine 2013); quantified population-based patient-reported barrier to care in Uganda, Kenya, Tanzania (CID 2016) and Zambia (CID 2019).

We draw from disciplines outside of traditional health sciences, such as marketing, to capture the patient (i.e., “customer”) preferences using discrete choice experiments. We found that patients are willing to incur considerable time and costs to access nicer services; (PLOS Medicine in 2018) as well as counter-intuitive preferences about service architecture.

We have started to use human centered design principles to inform service improvements.


We have been applying approaches we previously used in HIV to examine COVID-19. The Lorenz Curve, previously used to assess disparities in HIV viral load within a clinic as well as mortality across facilities. We now apply it to examine distribution and racial inequities in COVID-19 testing, as well as other analyses to examine disparities. Multi-state models are an excellent was to describe longitudinal courses when many “states” are temporary – previously applied to HIV treatment and now to understanding COVID-19 hospitalization course.

Presentations and Talks


2021 – 02.16 Fogatry – Adaptive Strategies for Engagement in HIV Care

2021 – 07.21 IAS 2021 Track D (Social, Behavioral and Implementation Science) Rapporteur Session

SIRC 2022 – Mechanisms

Mechanisms in implementation science: when the parts are greater than the sum

SIRC 2022 – Generalizing

Context as “King” of “Trap?” for Generalization: Mechanisms as a Middle Path

All Publications