@LancetGH examines the "collision" of #HIV and chronic disease (among other things)

Friday, January 26, 2018

The first thing that caught my eye in the February edition of Lancet Global Health was an article on the prevalence of COPD among persons living with HIV. The issue of care and treatment of non-communicable diseases among PLWH has become increasingly important as HIV itself has transitioned from a life-threatening condition to a chronic disease to be managed. However, what I found interesting about this particular meta-analysis was that its aim was to assess the hypothesis that HIV infection itself makes people more likely to develop COPD, even when controlling for tobacco use:
Results from several original observational studies and narrative reviews have suggested that the prevalence of chronic obstructive pulmonary disease (COPD) would be increased in people with HIV and even higher than in the general population. Findings from studies with similar designs also suggested an association between HIV infection and COPD.
...
Using strong and robust statistical methods, we found an association between HIV and COPD, even after adjustment for tobacco consumption, the leading risk factor for COPD. We have also shown a high prevalence of COPD in HIV-positive individuals. We identified three factors favouring an increase in the prevalence of COPD among HIV-positive people: tobacco consumption (a common factor between people with HIV and the general population), the presence of a detectable HIV viral load (in HIV-positive people only), and country level of income. We found that the prevalence of COPD in HIV-positive people increased with the level of income of the country.
Interestingly, the accompanying commentary also pointed out that, while the highest HIV burden is borne by LMICs, there were very few studies based on such countries that were robust enough to be included:
Although most of the included studies were done in high-income countries in Europe and North America, most of the people with COPD and HIV live in LMICs. Only four of the included studies were done in Africa, the WHO region with the most people living with HIV. This finding is similar to what we found in 2013, when we did a systematic review of COPD studies in sub-Saharan Africa and found only one high quality prevalence study. This Burden of Obstructive Lung Disease (BOLD) study was done in South Africa and found a high prevalence of COPD: 22% of men and 17% of women. Taken together, these observations highlight an important imbalance between where this kind of research is done and where the need is greatest.
The journal issue contains a number of other tasty morsels as well. I thought this piece on indigenous languages as a factor in access to care was particularly interesting, as well as this one on "equitable access" to (unpaid) WHO internships (and boy, do I have a soapbox about that). One contributor from Stanford even held up a mirror to the journal, analyzing the geographic focus and authorship of all of its articles to assess its progress toward its stated aim to represent “disadvantaged populations” in health-related scenarios around the world:
From all 236 articles, only about 35% (SD=0·31) of the authors were affiliated with or came from LMICs. Articles on Africa had 44% (SD=0·28) LMIC authorship, south Asia had 52% (SD=0·29), southeast Asia had 56% (SD=0·35), and the Middle East and north Africa had 28% (SD=0·33). The Americas had 33% (SD=0·41). Multiregional articles had 17% (0·24) LMIC authorship.

The apparent under-representation of authors from LMICs contributing to articles focusing on LMICs highlights several issues. These include so-called safari research that recruits LMIC specialists, with minimum involvement, into studies driven by high-income-country authors for perceived credibility, the scarcity of grants in LMICs, low awareness of fee waivers within open access journals, insufficient infrastructure for large-scale studies of high impact, and not enough funding directly to academic institutions in LMICs.
It is worth pointing out that even awareness of open access fee waivers does not make them easy to get. In my efforts to get a paper on the IH Section's global health jobs analysis published in BMC Public Health, I engaged in a snarky exchange with the waivers department to explain (repeatedly) that we were volunteers and not part of an academic institution - and didn't get the waiver until I publicly called them out on Twitter. I can't imagine being in a situation like that and trying to make my case in a language that is not my own.

However, my favorite bit by far came from a commentary piece on an outcomes assessment of a healthcare "social franchise" (whatever the heck that is) in Uttar Pradesh. The author throws some not-so-subtle shade toward aid-funded projects and their reticence to rigorously evaluate their own effectiveness:
More importantly, and perhaps inadvertently, the study has shed light on more fundamental questions: why has social franchising as a model expanded at an exponential rate when there was little rigorous evidence of the model's impact on population health? Why have millions of dollars, often taxpayer money, been poured into an unproven idea? Why is there a paucity of rigorous research in documenting effectiveness of this heavily invested idea?

Perhaps the last question might be the easiest to answer. [...] The remit of major bilateral donors is to provide services that will directly improve the health of recipient populations. They are not in the business of research. Consequently, most donors limit their data collection activities to before and after surveys under the rubric of evaluation. They tend to be averse to funding data collection from control sites where their programme was not implemented.

Additionally, bilateral donors, as custodians of taxpayer money, operate in environments characterised by structural disincentives to acknowledge when programme efforts are not achieving their intended results.
Burn.

Using satellite data to predict cholera outbreaks

Thursday, January 25, 2018

I came across a nifty piece in Scientific American this week about a group of scientists whose algorithm to predict inland cholera outbreaks was unexpectedly proven to be accurate:
Quickly collecting ground data about [cholera epidemics] can be challenging, especially in chaotic locations. Yemen is a textbook case. “Yemen has massive civil unrest, people are moving around, [there is] political instability—there’s no way for us to get a single data point,” Jutla says. But satellites gave his team a way to assess the disease risk from the sky, and without being in the country.

At the American Geophysical Union annual meeting in December, Jutla presented the group’s prediction model of cholera for Yemen. The team used a handful of satellites to monitor temperatures, water storage, precipitation and land around the country. By processing that information in algorithms they developed, the team predicted areas most at risk for an outbreak over the upcoming month.

Weeks later an epidemic occurred that closely resembled what the model had predicted. “It was something we did not expect,” Jutla says, because they had built the algorithms—and calibrated and validated them—on data from the Bengal Delta in southern Asia as well as parts of Africa.
It's important to keep in mind that just because the model got this instance correct does not mean there is not room for improvement. However, it seems to have support from other experts:
“One of the things I like,” says Michael Wimberly, an ecologist at South Dakota State University, is that they are not looking “only at correlation to rainfall.” ...He says the cholera model is well grounded in hydrology and epidemiology. “They have an understanding of different types of epidemics that occur in different seasons; it’s very sophisticated."

#OpenData Highlights: Air pollution in London

Wednesday, January 24, 2018

Last week's edition of Data is Plural featured a data set on air pollution in London:
London air pollution. The London Air Quality Network, run by researchers at King's College London, gathers data on levels of nitrogen dioxide, ozone, fine particulate matter, and other pollutants from more than 100 monitoring sites. You can download the data as CSV files (for up to six metric and site combinations at a time) or fetch JSON and XML data from the site’s API. Related: London air pollution live data – where will be first to break legal limits in 2018? ” (The Guardian). Previously: Air quality data from the EPA (DIP 2017.10.04), OpenAQ (DIP 2017.03.29), Berkeley Earth (DIP 2017.03.22), and the World Health Organization (DIP 2016.06.15).
Air quality data "across the pond" here at home was featured last year. I like that I get to see some repeat themes in these newsletters, including medical image scans, foodborne illness, and air and water quality.

Last week in @CDCMMWR: #HIV among PWID in the US and young women in Africa

Monday, January 15, 2018

This week's edition of MMWR features two analyses on HIV. The first is on infection prevalence and risk among persons who inject drugs in 20 U.S. cities. The data comes from the National HIV Behavioral Surveillance system:
In 2015, National HIV Behavioral Surveillance found a 7% prevalence of HIV infection among persons who inject drugs which was lower than in 2012 (11%). Among HIV-negative respondents, 27% reported sharing syringes and 67% reported having vaginal sex without a condom in the previous 12 months; only 52% received syringes from a syringe services program and 34% received all syringes from sterile sources. HIV infection prevalence was higher among blacks (11%) than whites (6%) but more white persons who inject drugs shared syringes (white: 39%; black: 17%) and injection equipment (white: 61%; black: 41%) in the previous 12 months.
What I find interesting is that the prevalence of unsafe injection practices is so much higher among whites than blacks, although this may be related to there being so many new white PWID due to the opioid crisis (i.e., more black PWID have been injecting for longer). What does not surprise me, sadly, is that so few PWID were able to access sterile syringes from a syringe exchange program (SEP). Even after Congress lifted the federal funding ban on SEPs, states and other jurisdictions have been reluctant to operate them.

The second article (which boasts a jaw-dropping 82 co-authors from 22 institutions) reports HIV status and treatment cascade metrics for women aged 15-24 in seven countries in eastern and southern Africa:
Analysis of data from Population-based HIV Impact Assessment surveys conducted during 2015–2017 in seven countries in Eastern and Southern Africa found that the prevalence of HIV infection among adolescent girls and young women was 3.6%. Among those who were HIV-positive, 46.3% reported being aware of their status, and among those aware of their HIV-positive status, 85.5% reported current antiretroviral treatment (ART) use. Overall, viral load suppression among HIV-infected adolescent girls and young women, regardless of status awareness or current use of ART, was 45.0%, well below the UNAIDS target of 73%.
While low levels of awareness of status and viral suppression are pretty depressing, I was encouraged to see that so many who are aware of their infection are on ART. I was also fascinated to learn that the PHIA survey used is funded by PEPFAR:
The PHIA surveys are nationally representative, household-based surveys funded by the U.S. President’s Emergency Plan for AIDS Relief (PEPFAR) and conducted under the leadership of the respective countries’ ministries of health, CDC, and ICAP at Columbia University). The objectives of the PHIA surveys are to provide national estimates of HIV incidence and subnational estimates of HIV prevalence and viral load suppression to assess the HIV epidemic and the impact of HIV prevention and ART programs in each country. During 2015–2017, PHIA surveys were conducted in Lesotho, Malawi, Swaziland, Uganda, Tanzania Zambia, and Zimbabwe.
Also super cool: the survey takers conduct HIV, CD4, and viral load testing on the spot:
The surveys included home-based HIV counseling and testing conducted in private locations within or around the home, using each country’s national HIV rapid testing algorithm, and employing CD4 testing technology, with results immediately returned to participants. Awareness of HIV status and current ART use (an indicator of ART coverage at the population level) were determined based on responses provided in the survey questionnaire. HIV viral load testing was conducted using plasma specimens or dried blood spots.

@CDCgov researchers report increases in HCV infection and #opioid admissions in @AMJPublicHealth

Friday, January 12, 2018

Just before Christmas, researchers from the CDC's Division of Viral Hepatitis published an article in the American Journal of Public Health reporting an increase in both acute hepatitis C (HCV) cases and admissions for injection drug use in substance abuse treatment centers:
The annual incidence rate of acute HCV infection increased more than 2-fold (from 0.3 to 0.7 cases/100 000) from 2004 to 2014, with significant increases among select demographic subgroups. Admissions for substance use disorder attributed to injection of heroin and prescription opioid analgesics increased significantly, with an almost 4-fold increase in prescription opioid analgesic injection. Significant increases in opioid injection mirrored those for reported cases of acute HCV infection among demographic subgroups.
CDC featured the article, along with several related graphics, in a press release. As a side note, HCV infection is considered a highly reliable proxy for injection drug use, although I would advise caution when looking at "acute" HCV figures. HCV infection is frequently asymptomatic (i.e., not acute), and surveillance for acute cases is spotty in some states. Overall HCV infection rates are most likely much higher, however, meaning that these numbers point to what is most likely a much larger overall infection rate.

The authors used national surveillance data for HCV infection and substance use treatment data from SAMHSA for IDU admissions:
We obtained confirmed cases of acute HCV infection and associated demographic and risk characteristics from the National Notifiable Disease Surveillance System (NNDSS) for 2004 to 2014.
...
TEDS is a national data system administered by SAMHSA. It collects information on annual admissions to SUD treatment facilities in the United States. TEDS contains data on admissions to publicly funded and state-certified SUD treatment facilities by year and by state of treatment facility for all persons aged 12 years or older. By state law, treatment facilities provide data to TEDS. TEDS is estimated to include 67% of all SUD treatment admissions and 83% of TEDS-eligible admissions in the United States.
Despite limitations of the data, this should add to the list of rather loud alarm bells that we have a serious injection drug use problem. HIV looms.

Interestingly, the same issue of AJPH featured a commentary on the national opioid crisis, co-authored by HIV/IDU heavyweight Daniel Ciccarone at UCF. After a brief historical overview that separates the crisis into three phases, the piece criticizes what it calls the "vector model" - the focus on supply of opioid prescription drugs as the root of the crisis - and argues that appropriate policy responses should instead consider the reasons behind the demand for such drugs. The authors present several related possiblities to explain demand, including "diseases of despair" and the structural aspects of poverty:
The “reversal of fortunes” in life expectancy saw rapid diffusion, going from largely limited to Appalachia and the Southwest in 2000 to nationwide by 2015. The unprecedented 20-year difference in life expectancy between the healthiest and least healthy counties is largely explained by socioeconomic factors correlated with race/ethnicity, behavioral and metabolic risk, and health care access. These indicators are the most recent evidence of a long-term process of decline: a multidecade rise in income inequality and economic shocks stemming from deindustrialization and social safety net cuts. The 2008 financial crisis along with austerity measures and other neoliberal policies have further eroded physical and mental well-being.
It's an excellent piece that looks at the data on structural factors behind other forms of substance use (e.g., alcoholism), issues of racism, and the inability of the current U.S. health care system to adequately address the problem. It urges a human-centered approach to the problem and says that we should "focus on suffering."

More #opendata highlights: @CDCgov's 500 Cities project!

Thursday, January 11, 2018

The 500 Cities project was featured in Data is Plural this week!
Local health metrics.The CDC’s 500 Cities Project provides “city and census tract-level data, obtained using small area estimation methods, for 27 chronic disease measures for the 500 largest American cities.” The metrics range from cancer prevalence to binge drinking to dental health to undersleeping. The latest data release was published in December and covers more than 28,000 Census tracts.
I'm irrationally excited about this initiative and the potential for discovery in the data. So much so that I mentioned it in an upcoming installment of an open data series I am working on for Cadence Group.

#Opendata highlights: Mammographies

Tuesday, January 9, 2018

The very last edition of Data is plural in 2017 featured an improved database of mammographies, overhauled by researchers at Stanford:
A better mammography database. The Digital Database for Screening Mammography was first released two decades ago, in 1997. It contains data and images from 2,620 mammographies — a mix of normal, benign, and malignant cases. In a Scientific Data article published last week, a team of Stanford University researchers describe a series of improvements they’ve made to the original database; their Curated Breast Imaging Subset of DDSM has modernized the database’s image formatting, added detailed “region-of-interest” annotations, and converted the metadata into CSV files.
This one hit home for me because my mother's sister goes in for surgery today for a lump they found in her mammary duct. Va com Deus, Tia.