Over the past 5 years, hollow fiber membrane microfilters have been introduced into much of the developing world to combat waterborne illness stemming from microbially compromised water sources. One such filter, the PointONE™ Filter (Sawyer Corporation) has performed well in laboratory trials, in the 5 and 6 log reductions of protozoan parasites and bacteria, respectively (Hydreion LLC 2005). In a field study conducted in Cochabamba, Bolivia and recently published by Lindquist et al. (2014), households using PointONE filters had significantly less diarrheal disease compared with the control arm during the intervention period. Diarrheal prevalence ratios of 0.21 (95% confidence interval [95% CI] = 0.15–0.30) were observed for the filter arm and 0.27 (95% CI = 0.22–0.34) were observed for the filter and WASH BCC (education) arm. These diarrheal disease reductions occurred in only a 3-month intervention period, and showed marked improvement in the health of children under the age of 5 years old. An in situ study on long-term filter performance has been an important need for organizations that are currently using or are considering using these filters in the developing world. The relatively new introduction of the PointONE filter has precluded long-term study of its performance before this time.

In this issue of Journal of Water, Sanitation and Hygiene for Development, Murray et al. (2015) endeavor to contribute to this needed body of information. In particular, their study suggests several potential shortcomings of the PointONE filter after household use over the course of nearly 2 years. These include: filter fouling, sediment buildup, discolored membrane fibers, and membrane rupture. Admittedly, if such shortcomings were verified through rigorous scientific study, they would certainly cast doubt on the long-term effectiveness of hollow fiber membrane microfilters for household level point of use in the developing world. A long-term field study of the PointONE filter is an admirable undertaking and has the potential to answer important questions on log-term filtration efficacy, filter longevity, and effective life cycle.

In review of this research article, we deem it necessary to bring to light several substantial concerns we encountered with respect to the methods used and the claims made. Shortcomings in this article can be seen in the following areas: (1) poor pre-analysis filter storage conditions; (2) crude filter cartridge entry; (3) small sample size; and (4) inconsistencies in the article figures.

POOR PRE-ANALYSIS FILTER STORAGE

In our view, the most significant oversight of this study is that, to the best of our understanding, the investigators took filters collected in the tropics (Honduras), sealed them in a plastic bag, and undertook the cleaning and analysis 2 months later. Microbiologically, these conditions would appear to promote microbial growth and thriving from the moment of sealing. If water from the input side of the filter were to have spilled into the inner surface of the storage bag, chances are good that they could have reached the output side of the filter. This potential for contamination is too great to be ignored. Likewise, if the transport of the bag-stored filters was in the cargo hold of an aircraft where freezing occurred, the water held in the pores of the hollow fiber membrane could very easily have expanded and applied a tearing force to the fibers. Detail of physical placement and ambient conditions during transportation was not provided. Likewise, insufficient detail was given as to whether all six of the filters were individually sealed in a bag, or collectively were combined into one bag. Had the latter scenario been followed, then the possibility of cross-contaminating input and output water is problematic to this study.

A better method would have been to have each filter cleaned (as per manufacturer's instructions), each side sealed to prevent input–output contamination, single-filter transportation bag storage for transport, then immediately analyzed upon arrival to the host institution laboratory (within 48 hours). However, ideally, the microbiological testing should have been done in situ, in a situation where contamination and storage- and transport-related methodologies would not introduce doubt into the methods used.

CRUDE FILTER CARTRIDGE ENTRY

In Figure 2, the crude manner of filter cartridge entry affects the visual interpretation of the photos (Murray et al. 2015). It is clearly visible that plastic fragments and powder from the membrane cartridge housing have fallen onto the input end of the filter fibers during the entry into the filter cartridge. We were left wondering how much of this minute plastic debris was depicted as the fouling layer in Figures 3 and 4.

SMALL SAMPLE SIZE

The interpretation of the results should recognize the uncertainty due to the small sample size (n = 6) and biased sample. It is unclear if the the six filters selected for evaluation were ones that showed poor results by Goeb (2013), which would not provide a representative sample of the whole. We have included the citation for Goeb (2013) here, but could not locate this article online or in any library resource in order to verify the sample collection methods.

INCONSISTENCIES IN THE ARTICLE FIGURES

This article mentions burst fibers, yet no photo is shown. In Figure 2(b), it does appear as if the entry method into the filter cartridge may have damaged some filter fibers (in the upper left of the image) (Murray et al. 2015). We were puzzled as to why the comparative images in Figure 3 did not use equal magnifications for comparison. In the cases where a flaky fouling layer is seen (Figure 3(e) and (f)), magnifications are much higher than the comparator (new filter). This appears misleading, especially if the flakes are minute residues of plastic from the filter cartridge entry method. Figure 4 did use equal magnifications for direct comparisons, yet we had difficulty seeing the fouling that the text discussed.

FINAL REMARKS

Other aspects of this article raised questions. In Lindquist et al. (2014), household caregivers were trained on filter usage and cleaning, then were later (2–4 weeks later) tested on these skills. We wondered if the same was done in the Trojes, Honduras communities; not just training. Lastly, the introduction cites many publications by Goeb which we could not locate online or in library resources. Whereas the Murray et al. (2015) microbiological methods are known, Goeb's are unknown.

VERIFYING RESULTS IN THE LONG TERM

Although this study raises some potentially important concerns for long-term use of hollow fiber membrane microfilters, many of these seem to be left unsubstantiated in large part due to the study methods selected by the authors. At the same time, this article does point to the need for a microbiological study in the near future on long-term in situ filter performance in communities that have been using these filters for extended periods. As such, it is our intention to conduct such a study for filters in use over a 5-year period, and it is our hope that the findings from this study would be welcomed for future publication in the Journal of Water, Sanitation and Hygiene for Development.

REFERENCES

Goeb
M.
2013
Follow-up on Sawyer Filters in the Community of San Francisco de las Quebradas, Trojes, Honduras
.
Pure Water for the World
,
Rutland, VT
.
Hydreion LLC
(
Missoula, MT
),
2005
Microbiological Testing of the Sawyer 7/6B Filter
.
Report No. S05-03. Available at: http://sawyer.com/wp-content/uploads/2013/12/field-micro.pdf. Accessed 29 December 2014
.
Lindquist
E. D.
George
C. M.
Perin
J.
Neiswender de Calani
K. L.
Norman
W. R.
Perry
H.
Davis
T. P.
2014
A Cluster Randomized Controlled Trial to Reduce Childhood Diarrhea Using Hollow Fiber Water Filter and/or Hygiene-Sanitation Educational Interventions
.
Am. J. Trop. Med. Hyg.
91
(
1
),
190
197
.
Murray
A.
Goeb
M.
Stewart
B.
Hopper
C.
Peck
J.
Meub
C.
Asatekin
A.
Lantagne
D
2015
Fouling in hollow fiber membrane microfilters used for household water treatment
.
Journal of Water, Sanitation and Hygiene for Development
5
(
2
),
220
228
.