Here's a good one I found by a guy who was stupid enough to fall for that Rockwell paper when hiking the PCT:
http://bucktrack.blogspot.com/2011/03/waterborne-giardia-for-backpackers-no.html

Years ago I twice contracted Giardiasis, in consecutive years. Since the second time I have been very careful about treating water, and had no trouble for many years. However, before I left on my five month thru-hike of the Pacific Crest Trail last year, I read this article. Giardia lamblia and Giardiasis With Particular Attention to the Sierra Nevada. It made a very convincing case that the odds of getting Giardia from backcountry drinking water was so low it didn't pay to worry about, at least in the High Sierras, as long as some attempt was made to select water sources. Instead, a focus on hygiene was stressed. So on the PCT I usually didn't treat my water. But I was the most careful I've ever been with hygiene in the outdoors. Unfortunately I got so sick in the Sierras that I was curled up on a mountainside, and it took me most of a day to make it the four miles (only four, luckily) to the nearest road. I partially recovered, after a full day of rest, then a week later became very sick again. Luckily I was at Mammoth Lakes. My physician diagnosed Giardia. He told me he treats many backpacker Giardia cases and doesn't report it. He laughed when I told him about the article I read. Other hikers have told me their doctors told them the same. A cursory scanning of Pacific Crest Trail sources shows about 12 thru-hikers a year (out of around 300) report getting giardiasis. And despite statements to the contrary, most are diagnosed by lab tests and/or physicians. With 2/3 of people being asymptomatic, the true number is likely much higher.

Edited by hhope on 09/16/2012 21:37:30 MDT.

Looking further at the abstract for the 1977 study in Colorado, I see some interesting numbers.

"A one-year retrospective laboratory survey in Colorado revealed that 691 (3%) of 22,743 stool examinations for ova and parasites were positive for Giardia lamblia, a higher percentage than that reported from surveys outside of Colorado."

That means that out of 100 people who had persistent stomach upset, went to the doctor, and convinced the doctor to order an ova and parasite test, only 3 of them were found to have giardia.

"a higher proportion of cases than controls who visited Colorado mountains (69% vs. 47%), camped out overnight (38% vs. 18%), and drank untreated mountain water (50% vs. 17%)"
So if 17% of the Colorado population of about 4 million in 1977 drank untreated water that would be 680,000 people. Out of these 691 tested positive for giardia, that is a 0.1% lab identified infection rate. This is assuming that all giardia cases in Colorado were caused by drinking untreated water, which of course is an overestimate.

So if you drank untreated water in Colorado in 1977 you had less than 1 in a 1000 chance of catching giardia so bad that you went to the doctor and got it lab tested and confirmed.

I'm a nerd but not a statistician, so if I am reading this wrong let me know.

Edited by KatyAnderson on 09/16/2012 21:36:36 MDT.

Katy said: A one-year retrospective laboratory survey in Colorado revealed that 691 (3%) of 22,743 stool examinations for ova and parasites were positive for Giardia lamblia, a higher percentage than that reported from surveys outside of Colorado."

That means that out of 100 people who had persistent stomach upset, went to the doctor, and convinced the doctor to order an ova and parasite test, only 3 of them were found to have giardia.

I don't think that can be true. 3% is, as they noted, "a higher percentage than that reported from surveys outside of Colorado" [who have giardia] so it must be cross section of the public in general. I believe something over 2% of people in the U.S. have giardiasis. Most backpackers I know who are tested for giardia come up positive.

Oh, coincidentally, I was the guy Harald quoted above who was stupid enough to believe the "giardia myth" papers. That is what triggered my research.

Edited by Colter on 09/17/2012 05:04:27 MDT.

> The "city water is dirtier than Sierra water" has been debunked.

Can you direct me to where this has been "debunked"? It's one of the three key pieces of evidence, and perhaps the most germane.

Thanks.

Bill S.

To quote what I've written earlier: A major theme in the Rockwell paper is that Sierra water on average is purer than San Francisco and LA water. Except for two big factors: these cities filter and/or treat their water supplies, so they are undoubtedly much safer than untreated Sierra water. And hikers don't drink an "average" canteen of water. Their water bottle might contain no Giardia cysts, or they might get enough cysts to send them to the hospital.

http://bucktrack.blogspot.com/2011/03/waterborne-giardia-for-backpackers-no.html

I would also add that the giardia numbers used for the Sierras are decades old and undependable.

Bruce,

In my opinion, if you want to refute the validity of the comparisons among Sierra and California city water supplies, you need to dig deeper into the data and show exactly where and how they're flawed. I have only looked at a few of the reports in detail, but if I have understood and recall them correctly, the Hetch Hetchy water feeding the Bay area was considered clean enough not to require filtration. If that's the case, then the count comparison per se may well be valid.

What you may want to look into is whether there's evidence that would bear upon the viability of those cysts. If I've understood correctly, the water does get chlorinated, and while the relevant chlorine concentrations would be inadequate for purifying a more highly contaminated source in the field, they may be enough - over an extended period of exposure - to knock the viable cyst count down significantly (i.e., to a level well below that of contaminated Sierra sources). Just to be clear, I'm speculating as to what you might find, but I think that's the place to be looking for clues.

The other place, as I've intimated before, is the original data on the number of cysts required for infection. Some oft-cited sources have taken the data to say that infection cannot occur below a 10 cyst exposure (or that the probably is dramatically lower at lower exposures), a clear mis-interpretation of the findings. The implications for the potential infectivity of water with low level contamination are huge. Finding ten in a given liter of water at typical measured Sierra levels is enormously improbable. Finding one is going to happen a lot more often.

And you are right, of course, that it's inappropriate to draw conclusions that assume that cysts are randomly distributed in Sierra water. This profoundly alters the odds that have been calculated based on measured concentrations (interestingly, unless one cyst is a tenth as likely to infect you as ten cysts).

Bill S.

At some point enough is enough on digging too deep in a blog post. The typical person is unlikely to read all the way through a long post with citations as it is. If it were a paper for peer review, you'd probably be right.

I know Hetch Hetchy water is run through water treatment plants and is treated with Chloramine. I found this quote: "Giardia cysts can be inactivated by free chlorine at 0 to 25 degrees Celsius, or with carbon dioxide, ozone, or chloramine." (What's In the Water? Climate Change, Waterborne Pathogens, and the Safety of the Rural Alaskan Water Supply) The Rockwell paper failed to mention treatment plants at all. One skeptic, a friend of Rockwell, pointed out it was because at the time they were using chlorine and chlorine is ineffective. If that's the truth, Rockwell was wrong on that as well:Due to a misstatement in the early literature, Giardia cysts have been considered for many years to be extremely resistant to disinfection...Several studies have shown that as little as 2 mg of chlorine per liter can kill greater than 99.8% of the Giardia cysts

http://www.tandfonline.com/doi/abs/10.1080/10643388809388341

I think we can agree that the whole city/Sierra concentrion argument is invalid from the get-go anyway because high concentrations of viable giardia cysts are dramatically more likely in the Sierras compared to a city hundreds of miles away, especially after it's been run through a water treatment plant.

Sorry to drill down on this, but if we are to discuss science we have to look carefully at the numbers in these studies and try to understand what they mean.

Katy said: (the quote is from the 1977 study abstract)
"A one-year retrospective laboratory survey in Colorado revealed that 691 (3%) of 22,743 stool examinations for ova and parasites were positive for Giardia lamblia, a higher percentage than that reported from surveys outside of Colorado."

That means that out of 100 people who had persistent stomach upset, went to the doctor, and convinced the doctor to order an ova and parasite test, only 3 of them were found to have giardia.

Bruce said:
I don't think that can be true.

Katy says:
True or not, what this study found was that 3% of the stool exams were positive.

Bruce said:
3% is, as they noted, "a higher percentage than that reported from surveys outside of Colorado" [who have giardia] so it must be cross section of the public in general.

Katy says:
If it was a cross section of the public that would mean that 22,743 healthy people in Colorado had a ova and parasite stool test done for research purposes in 1977. Seems unlikely.
Instead I believe "a retrospective laboratory survey" is a study where they do a statistical analysis of existing lab results.
So these ova and parasite stool exams were done on patients rather than healthy research subjects. Yet only 3% showed up positive for giardia.

Hi Katy,

It's good that you don't just accept things as they appear to be on the surface.

One thing I've learned early on in this research is to step back and ask "here's what the numbers seem to show. Does this make sense?"

I don't doubt the 3% number. I doubt highly that it was 3% of people that had symptoms that physicians believed warranted a giardiasis test, if that is what you are saying. The overall incidence of infection in the United States is estimated at 2% of the population (FDA.) Colorado has a reputation for a higher than average infection rate. I think the ability of a physician to diagnose giardiasis is much better than random chance.

Perhaps the tests were from people with unidentifiable stomach ailments. I don't know. Too bad there isn't free public access to the full report.

Thanks for the skepticism!

Does it matter if people are infected with no symptoms? They might pass it on to other people or animals, but we should already treat fec es as though it's infected with whatever. One theory says we're too hygenic - if we are exposed to some bugs we're less likely to have auto-immune diseases.

Another question is, does treating water prevent you from getting infected? maybe with or without symptoms?

Like, people say that they got Giardiosis from swimming - but that's irrelevant to the question of treating water. Or people have said they got Giardiosis even though they treated their water - maybe that would indicate that it's hopeless so don't worry about it or you have to do it more carefully.

A good study would be to select 100 people at a trailhead, randomly either select no water treatment or either of several treatments, maybe give them good instruction how to properly treat, test them before and after a backpack trip. You'de still have a bias because only some people would agree to it.

"A good study would be to select 100 people at a trailhead"

Most studies get better results if they are blind studies. In other words, the testees can't know exactly what method they are using. In the case of water treatment, I don't see how you could do that very easily.

Now, if you had three different chemicals that you wanted to compare, you could give one chemical to members of different groups, and they would not know exactly which one they got. Then you would have a control group that got nothing or a placebo, and you might have another control group that was using something completely different like a filter.

I don't want to be in a group that gets nothing.

--B.G.--

Bruce said:
Too bad there isn't free public access to the full report.

Katy says:
Agreed, without the full report of the Colorado study it is hard to draw any conclusions from these numbers.
Who exactly got the o&p stool test? Healthy study subjects or sick patients?
How was the control group selected?
Your assumption was that they divided them into two groups dependent on whether or not they treated their water and then reported giardia infection numbers for each group.
My assumption was that they compared the giardia infected subjects to a healthy control group, and reported behavior numbers (drank untreated, camped overnight, visited Colorado mountains) for each group.
Depending on which assumption is correct yields a very different reading of this study.

Bruce said:
Thanks for the skepticism!

Katy says:
Thanks for bringing this topic up for discussion in a way that tries to analyze the underlying science rather than rehash everybodies opinions.
And especially thanks for putting up with me taking apart every number and data point in order to come to the bottom of the science.

Katy said: So if 17% of the Colorado population of about 4 million in 1977 drank untreated water that would be 680,000 people. Out of these 691 tested positive for giardia, that is a 0.1% lab identified infection rate. This is assuming that all giardia cases in Colorado were caused by drinking untreated water, which of course is an overestimate.

So if you drank untreated water in Colorado in 1977 you had less than 1 in a 1000 chance of catching giardia so bad that you went to the doctor and got it lab tested and confirmed.

I think we need to read the full report! These are all good arguments making me think, for sure.

Assuming your calculations are true, another way to look at those above numbers would be that over an 80 year lifetime you'd have an 8% chance of catching giardiasis so bad you'd be sick for a month. Several times the risk for a person who spends a lot of time in backcountry. If a typical Coloradan drinks untreated water one day a year, which seems high, actually, the typical thru-hiker might drink untreated water 100+ days a year if they aren't treating. All things being equal, they might go from 1 out of a thousand to 1 out of 10 chances of getting serious giardiasis that year. All fuzzy numbers, of course.

Like you say, though, a significant proportion of cases aren't caused by drinking bad water at all. On the other hand, many bad cases go undiagnosed and run their course. Many cases are treated empirically and never lab tested. Most cases are asymptomatic. People with giardiasis often come up with a negative test. And as I found out in my research, nearly 99% of cases go unreported. I'm not sure how that would affect the Colorado study. I do know the author's concluded These results indicated that G. lamblia is endemic in Colorado and that drinking untreated mountain water is an important cause of endemic infection.

Edited by Colter on 09/18/2012 04:23:08 MDT.

The study was based on 1973 population of 2.1 million. From the study, half of identified infections were in people who didn't report drinking untreated water. Those two factors offset - as a thumbnail calculation, the water drinkers had about a 1/1000 chance of winding up with diagnosed giardiasis in a given year. Backpackers' risk will be higher, since they almost certainly drink more often than the average person, but the one day/year estimate above for the average untreated water drinker is too low. Assuming accurate reporting, the 17% all drank at least once - in the absence of data, what the distribution and its average would be is anyone's guess, but my WAG would be closer to 10 than 1.

FWIW, you can push the numbers a bit farther, if you wish, to get a rough estimate that untreated water-drinkers were about 5 x more likely to contract diagnosed giardia than non-drinkers.

The definitions of cases and controls were as expected per correspondence above:

"A case was defined as any person having a positive stool specimen for G. lamblia during the study year. A questionnaire was administered by telephone to each case and to controls matched by age, race, sex, and place of residence."

Best,

Bill S.

Bill said:
FWIW, you can push the numbers a bit farther, if you wish, to get a rough estimate that untreated water-drinkers were about 5 x more likely to contract diagnosed giardia than non-drinkers.

Katy says:
Interesting. Can you walk me through that calculation?

Katy says: Interesting. Can you walk me through that calculation?

Here's one of the ways to get there: about 1/6 of the population drinks untreated water, 5/6 does not. There are 1/5 as many drinkers as non-drinkers, but they account for equal #s of cases (the 50-50 split among affected individuals). For that to happen, drinkers have to be ~ 5 x as likely to be affected. Make sense?

Best,

Bill S.

Bill said: "A case was defined as any person having a positive stool specimen for G. lamblia during the study year. A questionnaire was administered by telephone to each case and to controls matched by age, race, sex, and place of residence."

Good stuff Bill, where did you get the above information?

It looks like you were right about the controls, Katy. And likely about the original pool of people having already been tested before the study.

Bill said: "Here's one of the ways to get there: about 1/6 of the population drinks untreated water, 5/6 does not. There are 1/5 as many drinkers as non-drinkers, but they account for equal #s of cases (the 50-50 split among affected individuals). For that to happen, drinkers have to be ~ 5 x as likely to be affected."

Very interesting. I'm pretty sure I would never have looked at it that way!

I would like to note that so far we have escaped Godwin's Law. http://en.wikipedia.org/wiki/Godwin's_law

Does just mentioning "Goodwin's Law" mean that we have now made a Nazi reference so the discussion is ended?

and since you brought it up, you lost?

or does my mentioning Nazis mean that it's ended and I've lost?

Thanks Bill, your explanation makes perfect sense.

Bruce now you can go back and edit your blog post:
"Welch makes an even more egregious "mistake" in Giardiasis from Wilderness Water. Citing Giardiasis in Colorado: an epidemiologic study he says Responses indicated that 38% of cases vs 18% of controls had camped overnight in backcountry areas. He ignores the very next line in the abstract to that paper which says and drank untreated mountain water (50% vs. 17%.) There it is, in black and white. The infection rate was TRIPLE for drinking untreated mountain water in this large group. I think it's a great example of confirmation bias and is bad science, at best."

The correct reading of this data is that there is FIVE TIMES the infection rate.
So if you drank untreated you had 1 in 1000 rate of lab diagnosed Giardia infection in any given year, whereas if you treated your water you would have 1 in 5000 rate of lab diagnosed Giardia infection in any given year.

But if you had only a 1 in 1000 chance of being infected in one year, it's hardly worth thinking about.

And if you got unlucky, it's obnoxious but surviveable. And I'll normally be back home.

I keep going back and forth whether I'll bother treating water.