I’ve recently become convinced that effective malaria control in sub-Saharan Africa will require more than just bed nets and effective medicine. While these interventions will surely help, environmental modification to reduce the presence or virulence of mosquitoes, is essential. This is how malaria was eliminated in most countries where it has been successfully controlled, and I suspect indoor residual spraying (IRS) efforts have been key in many of the sub-Saharan African countries where potential success stories are brewing. These efforts get less attention from donors and researchers, perhaps because they are less visible, but also because spraying has accumulated its own set of baggage.

Spraying with insecticides is believed to adversely affect the environment. Anyone who knows anything about this will immediately think of DDT and Rachel Carson. While this is clearly a dark part of malaria control history, I gather that great efforts have been made to develop new insecticides that do not have the same environmental risks as some of the earlier generation products.

But spraying has an additional problem in that mosquitoes will develop resistance to the insecticide weakening their effectiveness over time. This is just one of those things, just like it is hard to develop antibiotics that don’t promote resistance in bacteria and drugs that don’t promote resistance for viruses. Where there are millions of them, there will always be a few that will figure out how to beat whatever we throw at it. Evolution always wins.

That is why a new idea by Andrew Read, Penelope Lynch, and Matthew Thomas is so interesting. They suggest that they can develop a “evolution-proof” insecticide for malaria control by realizing that it is really only older mosquitoes that transmit disease, usually well past the time when they have reproduced, so if we can target mosquitoes with an insecticide during the last phases of their life, than we can prevent transmission without inducing evolutionary pressure on the mosquito population. Kind of like the reverse logic of why late onset diseases like Huntington are hard to weed with evolutionary pressures.

Their proposition is based mainly on a modeling exercise that has shown that such an idea is feasible rather than promoting a product actually achieves this goal. But it does change the paradigm in our thinking about resistance and could pave the way to new products.

Thanks to Jack at Amherst for making me aware of this work in the economist.

Share on Facebook
 

Years ago before deciding what I wanted to work on for my dissertation, I remember speaking to a colleague who suggested that there still had not been great empirical work done on user fees and that the question of what impact cost sharing had on utilization of health services was still very much up for debate. At the time, I wondered why on earth he thought this was still an interesting question. Of course user fees reduce utilization of health services: that is econ 101, right? User fees were just a bad policy dreamt up by the World Bank and the IMF, something I learned from my masters studies. In my mind, the Bamako Initiative was up there with the marketing of baby formula and Pfizer’s testing of drugs in Nigeria on the world’s list of really-bad-global-health-stories. The truth is that I just did not know enough.

As it turns out, years later I became interested in the question of user fees through some research I was doing in Ghana. Ghana like many countries in Africa was struggling to increase health service coverage and also like in many other countries user fees had become highly politicized. Some countries had experimented with eliminating user fees or to exempt certain people or certain services from user fees. What was surprising, was that the empirical literature that existed did not always show what we would expect it to show: increase coverage. Some user fee introductions were associated with higher coverage. Some elimination of user fees did not show higher coverage. Some of these experiments actually failed and countries were forced to reintroduce user fees. In many cases the exemptions that were introduced became non-functional over time.

Part of the problem, as it always is in these cases, was methodological. Most policies were implemented at the national level making it hard to find a good control group and the general availability of data was so poor that it was hard to properly measure the effect of the policies.

But part of it also has to do what many economists or others who ignore the political economy of user fees fail to recognize: user fees don’t just reduce patient demand they also provide important incentives to health care providers. Since these two effects work in opposite directions, whether eliminating or implementing user fees will increase or decrease utilization is an empirical question. Randomized experiments on the effects of user fees on the demand for health services where the supply side is purposely well controlled by the investigators neglect this offsetting effect and therefore don’t translate well to the real world.

For a brief period starting in 2003, Ghana exempted women from paying delivery fees. The policy was rolled out to 4 of the regions before being implemented nationwide 20 months later. I have recently finished a working paper that looks at the effect of this policy on the utilization of services. I find what both the advocates and opponents of user fees would argue: eliminating user fees did increase the utilization of services, the policy may have adversely affected quality, and because the government ignored the incentives to providers, the policy stopped being implemented over time. Everyone was right.

The real question then is not whether or not user fees limit demand, it does as we would expect, the question is how to implement a policy that can target services to the groups we are interested in while preserving the provider incentives. Pre-paid heath insurance plans, such as the plan currently being implemented in Ghana is likely one answer. Voucher systems might be others. The full effect of such policies needs to be considered.

Comments and criticisms of the paper, as it is still an early stage working paper, are very much appreciated.

Share on Facebook
 

I just was alerted to some really interesting research reported on by NYU Professor Marion Nestle in the Atlantic via one of my favorite twitterers (is that a word?) and bloggers – Mark Bittman.

The investigators conducted a randomized trial of an intervention to increase water consumption among school aged children in Germany. Schools were randomly assigned to receive a new water fountain, each child received their own water bottle, and educational programs were introduce to promote water consumption. Almost a year later, the children in the intervention arm were no fatter than a year before but children in the control schools were more likely to report being overweight. In total, the risk of being overweight was reduced by a whopping 31% (!). Children in the intervention arm reported drinking more water, but interestingly no difference in the consumption of other beverages, it appears that the water was in addition to other beverages. Huh?

Really fascinating. I would not believe it had the experimental design not seem to be so clean. Perhaps the sample was small, which could explain these results as the finding was just statistically significant, or perhaps it really worked. Excuse me while I go and get a glass of water.

Share on Facebook
 

The other day, while watching TV I noticed an advertisement on TV encouraging teenagers to speak up in situations when their peers are driving recklessly to reduce accidents. The ad was a bit corny, but for whatever reason it was memorable. Would it have changed by behavior in any way when I was a teenager? Would it affect me now (memories of one relatively recent trip driving with a friend to Vermont going about 140 miles an hour -the friend was driving – in a snow storm at midnight comes to mind)?

My colleague Billy Jack and his co-author James Habyarimana have released an updated working paper looking exactly at whether such interventions are effective at improving the driving of Matatu drivers in Kenya. The basic idea of the paper is that bad driving has negative consequences on the rest of society (injuries to those in the van or those injured by the van, direct economic effects of the accidents, etc) and randomizes a VERY simple and inexpensive intervention aimed at reducing this behavior. The intervention was the installation of stickers telling Matatu passengers to speak up when their driver is driving recklessness. The intervention reduce the number of reported accidents, specifically reported accidents where drivers were at fault, greatly improving welfare with a simple and cheap intervention. Drivers in the treated buses reported significantly increased complaints from passengers suggesting that it was the behavior of passengers that influenced the driving of the drivers.

I guess my skepticism does not apply to such a situation. It seems that information and empowerment can have a significant impact on behavior in this context. It also suggests that similar types of low-cost targeted interventions may also be effective ways at reducing the burden of road traffic injuries around the world.

Share on Facebook
 

Burkinabes and bike helmets

On April 6, 2009, in Burkina Faso, road traffic injuries, by Karen Grepin


(Photo credit Helge Fahrnberger)

A few years ago, prior to starting graduate school (OK..it was a bit more than a few years ago, I’ve been in grad school for a very long time) I spent the better part of half of a year in Ouagadougou, Burkina Faso volunteering with a community-based HIV organization. While living there I was acquainted, directly or indirectly, with about 7 people who died during my 5-6 months there. I probably knew only about a hundred people directly so this was a relatively high mortality rate. Despite the fact that I knew a lot of people with HIV, most of whom were not accessing treatment (this was in 2000), only 1 of the 7 died from anything related to the epidemic. 5 died in a single car accident traveling to a family event in Bobo and 1 died from a motorcycle accident.

Motorcycles – or mobylettes – are king in Ouaga. I spent nearly ever day driving on the back of one. When I returned to Burkina in 2005 for some work on onchocerciasis I promised my husband that I would not get on another one. They are terribly dangerous things accidents are common daily events.

I guess that is why I really related to this news story about how, despite government efforts, uptake of bicycle helmets has been really slow in the country. I think the article did a good job at raising what might be some of the factors that might be influencing this slow uptake, poor enforcement, high cost of helmets, current practice, and credit constraints. Experiences in Nigeria, a relatively more prosperous country, at trying to enforce the use of helmets was largely unsuccessful. How wealthy does a country have to be before it begins to adopt more life saving behaviors? What types of interventions are likely to be the most successful?

Share on Facebook
 

When to start HIV treatment – part II?

On April 6, 2009, in HIV/AIDS, research, by Karen Grepin

There has been a long standing debate in the clinical literature about the optimal time to commence ARV therapy. I have blogged about this debate before. It is something I know a little about since it was related to my hono(u)rs thesis in undergrad (what we Canadians call “college” because we go to University not college) back in the late 1990s.

In the absence of a randomized trial, this debate has raged on for over a decade. Although guidelines have generally supported the view that earlier is better, concerns over resistance and side-effects led to treatment being initiated on average much later in many cases. In resource poor areas, guidelines have generally adopted the view that ARVs should be prioritized to those who are the sickest – a reasonable value judgement in the context of more limited financial resources. But how likely is it that these treatment guidelines do the most to save lives?

But say it was survival that we really cared about than what should we do? A recent large prospective cohort study conducted since the mid-1990s in the US and Canada and recently published in the New England Journal of Medicine finds that earlier the better when it comes to survival – much better in fact. Comparing patients that initiate treatment at higher CD4 blood cell counts to those with lower levels the study found very large differences in survival probabilities. Initiating at CD4 counts above 500 was associated with nearly a doubling of the survival probability to those that deferred treatment. Initiating above 350 vs. lower also had significantly higher survival probabilities. Given the large sample size of this study, they were able to control for factors that are believe to be predictors for mortality. This is certainly no substitute for conducting a randomized trial, but it does improve the generalizability of these findings.

While it goes without saying that because these findings are based on observational data, they should be interpreted cautiously. In terms of the biases, it seems reasonable to think that the healthiest patients may be those who would be likely to defer treatment at any given CD4 count level, which would only strengthen these findings. Regardless, these findings do seriously question current ARV treatment guidelines, in particular in resource poor settings.

Would ARV resources most appropriately allocated to patients with higher CD4 counts not the reverse? Should we care about cost-effetiveness or just survival?

By the way, Kudos to the New England Journal for making some of their top articles free to readers.

Share on Facebook
 

My good – and newly married – friend Sara Bleich recently published, along with her co-authors Emre Ozaltin and Chris Murray, a paper that explores the determinants of health care satisfaction using data from the World Health Survey. Basically they find that patient satisfaction and other factors are predictors of overall satisfaction but together actually explain only a small fraction of the variation in satisfaction reported by patients across European countries.

In recent years the concept of “health system responsiveness” has been developed and advanced by the World Health Organization to capture some of the aspects of health systems performance above and beyond the more objective measures related to health outcomes. It refers to “… the manner and environment in which people are treated when they seek health care” and it is subjective by nature. While we think it is important, enough so that it receives a great deal of emphasis in the overall health system performance exercises, but yet we know little of what drives it or even really agree how to define it.

The reasons for focusing on European countries exclusively was purposeful it order to assume that there were not major differences in health outcomes, which is debatable, but perhaps a reasonable approach. Although most of the countries in the sample have nearly universal access to health services, modes of financing do vary among the countries.

What I found to be the most interesting was that although there were not major differences in health outcomes, and people generally had access, there were still relatively large differences in overall measures of satisfaction among countries. What explains the fact that only 10% of Spaniards were very satisfied with their health system while over 70% of Austrian were. Culture? History? Expectations? How well their soccer team did at the world cup? National wine consumption? Who knows. If we don’t know, then to what extent are the factors that drive these differences factors that can be influenced by public policy? If not, to what extent should health system performance be based on these types of measures? Lots of unanswered questions.

Share on Facebook
 

1. Will pneumonia be the next disease du jour?

2. The third round of the Gates Foundation Grand Challenges Exploration program have been announced.

3. Lance Armstrong crashed so he has time to work on his new website. My travel plans for July are on hold.

Share on Facebook
 

Analytics Plugin created by Web Hosting