Small water supplies under the microscope: a case study on water quality and health from rural Alabama

Joe Brown PhD PE

Lecturer, Department of Disease Control, Faculty of Infectious & Tropical Diseases, London School of Hygiene & Tropical Medicine, University of Londonjoe.brown@lshtm.ac.uk

AcknowledgementsFunding: US EPA (STAR)Students: Olivia Johns and Maxwell Izenberg (LSHTM)Study participants, AlabamaPauline Johnson and group (UA)Julie Olson and group (UA)Christine Stauber (GSU)Christine Moe and group (Emory)Rick Gelting and group (CDC)Shadi Eskaf (UNC)Engineers Without Borders (UA)HERO & Pam DorrSamory Pruitt and CCBP (UA)

Alabama

Gasteyer and Vaswani 2004 From 2000 census data

Southeast’s $50 Billion drinking water capital needs estimate

Source: EPA 2007, Drinking Water Infrastructure Needs Survey Assessment, Shadi Eskaf EFC-UNC.

Alabama Florida Georgia Kentucky Mississippi North Carolina

South Carolina

Tennessee $ 0

$ 2,000

$ 4,000

$ 6,000

$ 8,000

$ 10,000

$ 12,000

$ 14,000

$ 0

$ 500

$ 1,000

$ 1,500

$ 2,000

$ 2,500

$ 3,000

$ 3,500

Drinking Water 20-Year Capital Needs Estimates in 2007 (billions of $)

Billi

ons o

f $

$ / S

ervi

ce C

onne

ction

Rural water supplies• More than 94 percent of the

nation’s 156,000 public water systems serve <3,300 persons

• Disproportionate share of outbreaks related to microbial contamination

• Face a variety of challenges:– Wide service areas and

disperse populations – O & M challenges– Meeting regulatory

requirements

Southeast’s 8,700 community water systems serve 58.5 Million People

Alabama Florida Georgia Kentucky Mississippi North Carolina

South Carolina

Tennessee 0

500

1,000

1,500

2,000

2,500

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%Large Systems Serving >10,000 People

Small Systems Serving <10,000 People

Percent of Population Served by Small Systems

Num

ber o

f Cim

mun

ity W

ater

Sys

tem

s

Source: EPA’s 2011 SDWIS data analyzed by the Environmental Finance Center at the University of North Carolina, Shadi Eskaf

PWS Violations Alabama 1997-2012Type of systems Population

servedViolations: HAA

Violations: TTHMs

Violations: coliform

Total violations

Total violations per 100,000 people

Large (10,001-100,000) 4,136,225 15 28 100 143 3 Top third per capita income 3,248,710 11 15 51 77 2 Middle third per capita income 664,795 4 5 21 30 5 Lower third per-capita income 222,720 0 8 28 36 16Medium (3,301-10,000) 1,027,417 38 69 160 267 26 Top third per capita income 408,615 10 25 50 85 21 Middle third per capita income 384,255 23 23 58 104 27 Lower third per-capita income 234,547 5 21 52 78 33Small (501-3,300) 362,352 65 128 273 466 129 Top third per capita income 124,756 13 25 70 108 87 Middle third per capita income 139,053 34 74 107 215 155 Lower third per-capita income 98,543 18 29 96 143 145Very Small (<500) 11,168 11 13 42 66 591 Top third per capita income 2,547 6 1 9 16 628 Middle third per capita income 3,928 4 9 10 23 586 Lower third per-capita income 4,693 1 3 23 27 575Total 5,537,162 129 238 575 942 17

> 8x

197x

L Talebi

> 40x

Alabama’s Black Belt region

• Common demographic and socio-economic characteristics– High poverty– High unemployment– Decreasing population– High percentage of minorities

• Common themes– Decaying infrastructure– Poor access to basic services

and health care– High percentage of vulnerable

people (the young, elderly and infirm, HIV+)

Results from scoping assessment• In collaboration with HERO• Customer complaints: outages, cloudy or

smelly water • High cost to connect and tariffs

– Post-meter leakages– Services disconnected

• Alternatives are potentially unsafe water sources (i.e. shallow wells near failing septic systems)

• Poor septic access and apparent function

Phase 1: pilot study

• Pilot cross-sectional study of 300 households in Hale County to examine:

• (i) drinking water quality at the household level (both private wells and county public supply),

• (ii) possible associations between water infrastructure characteristics and drinking water quality, and

• (iii) risk of Highly Credible Gastrointestinal Illness (HCGI)

Data collection overview• One survey on water use, basic demographics,

health, and access to and perception of infrastructure

• All households submitted a water quality sample at the time of the visit– Some households submitted two samples

• Analysis for fecal coliform, turbidity, pH, and total and free chlorine– Membrane filtration, EPA 1904/Standard Methods– pH and Cl measured at point of sampling, colorimetric

tests– Turbidity

Not in the Job description

Results overview

• 8% of system samples positive for FC• 30% of well samples positive for FC• No Cl in >35% of samples from the county water

supply system – Cl presence associated with FC

• Frequently intermittent service: 8% of households• 18% of households had no onsite sanitation• Wells more likely to be contaminated

Results overview, continued

• 12 cases of GI illness among 507 people: 2.4% 7-day prevalence– High risk

• Individuals whose drinking water was found to contain ≥ 1 cfu/100 ml of FC were 4 times as likely to have also reported HCGI in the previous 7 days as those whose water sample was negative for FC (<1 cfu/100 ml) – (OR 4.0, 95% CI 1.2 – 14)

Fecal Coliforms per 100 ml sample<1 1-10 11-100 101 +

All Households

260 27 12 1

County supply

206 12 5 0

Well 54 15 7 1

   FC <1 cfu/100ml 

 FC ≥ 1 cfu/100ml

 OR (95% CI)

 p-value

Number of households 263 42     Connected to system 182 16 0.27(0.14-0.54) <0.001 Well users 46 20 4.3 (2.2-8.5) <0.001 On-site septic tank

 172

 28

 0.51 (0.064 - 4.1)

 0.528

No or unknown sanitation 184 29 0.98 (0.47-1.9) 0.905 Reported “poor or very poor” Taste Odor Color Clarity Perceived safety

  2120231812

  31311

  0.89 (0.25-3.1)0.30 (0.039-2.3)0.80 (0.23-2.8)0.33 (0.043-2.6)0.51 (0.065-4.0)

  0.8510.2420.7300.2900.523

 Intermittent service (piped supply only)

 23

 2 

 0.52 (0.12 - 2.3)

 0.390

Functional problems, well 6 4 

4.5 (1.2 - 17) 0.024

 Free chlorine between 0.2 and 2 mg/l

 146

 7

 0.16 (0.069 - 0.37)

 <0.001

Mean free chlorine (mg/l) 0.691 0.164    

   FC <1 cfu/100ml 

 FC ≥ 1 cfu/100ml

 OR (95% CI)

 p-value

Number of households 263 42     Connected to system 182 16 0.27(0.14-0.54) <0.001 Well users 46 20 4.3 (2.2-8.5) <0.001 On-site septic tank

 172

 28

 0.51 (0.064 - 4.1)

 0.528

No or unknown sanitation 184 29 0.98 (0.47-1.9) 0.905 Reported “poor or very poor” Taste Odor Color Clarity Perceived safety

  2120231812

  31311

  0.89 (0.25-3.1)0.30 (0.039-2.3)0.80 (0.23-2.8)0.33 (0.043-2.6)0.51 (0.065-4.0)

  0.8510.2420.7300.2900.523

 Intermittent service (piped supply only)

 23

 2 

 0.52 (0.12 - 2.3)

 0.390

Functional problems, well 6 4 

4.5 (1.2 - 17) 0.024

 Free chlorine between 0.2 and 2 mg/l

 146

 7

 0.16 (0.069 - 0.37)

 <0.001

Mean free chlorine (mg/l) 0.691 0.164    

   FC <1 cfu/100ml 

 FC ≥ 1 cfu/100ml

 OR (95% CI)

 p-value

Number of households 263 42     Connected to system 182 16 0.27(0.14-0.54) <0.001 Well users 46 20 4.3 (2.2-8.5) <0.001 On-site septic tank

 172

 28

 0.51 (0.064 - 4.1)

 0.528

No or unknown sanitation 184 29 0.98 (0.47-1.9) 0.905 Reported “poor or very poor” Taste Odor Color Clarity Perceived safety

  2120231812

  31311

  0.89 (0.25-3.1)0.30 (0.039-2.3)0.80 (0.23-2.8)0.33 (0.043-2.6)0.51 (0.065-4.0)

  0.8510.2420.7300.2900.523

 Intermittent service (piped supply only)

 23

 2 

 0.52 (0.12 - 2.3)

 0.390

Functional problems, well 6 4 

4.5 (1.2 - 17) 0.024

 Free chlorine between 0.2 and 2 mg/l

 146

 7

 0.16 (0.069 - 0.37)

 <0.001

Mean free chlorine (mg/l) 0.691 0.164    

   FC <1 cfu/100ml 

 FC ≥ 1 cfu/100ml

 OR (95% CI)

 p-value

Number of households 263 42     Connected to system 182 16 0.27(0.14-0.54) <0.001 Well users 46 20 4.3 (2.2-8.5) <0.001 On-site septic tank

 172

 28

 0.51 (0.064 - 4.1)

 0.528

No or unknown sanitation 184 29 0.98 (0.47-1.9) 0.905 Reported “poor or very poor” Taste Odor Color Clarity Perceived safety

  2120231812

  31311

  0.89 (0.25-3.1)0.30 (0.039-2.3)0.80 (0.23-2.8)0.33 (0.043-2.6)0.51 (0.065-4.0)

  0.8510.2420.7300.2900.523

 Intermittent service (piped supply only)

 23

 2 

 0.52 (0.12 - 2.3)

 0.390

Functional problems, well 6 4 

4.5 (1.2 - 17) 0.024

 Free chlorine between 0.2 and 2 mg/l

 146

 7

 0.16 (0.069 - 0.37)

 <0.001

Mean free chlorine (mg/l) 0.691 0.164    

   FC <1 cfu/100ml 

 FC ≥ 1 cfu/100ml

 OR (95% CI)

 p-value

Number of households 263 42     Connected to system 182 16 0.27(0.14-0.54) <0.001 Well users 46 20 4.3 (2.2-8.5) <0.001 On-site septic tank

 172

 28

 0.51 (0.064 - 4.1)

 0.528

No or unknown sanitation 184 29 0.98 (0.47-1.9) 0.905 Reported “poor or very poor” Taste Odor Color Clarity Perceived safety

  2120231812

  31311

  0.89 (0.25-3.1)0.30 (0.039-2.3)0.80 (0.23-2.8)0.33 (0.043-2.6)0.51 (0.065-4.0)

  0.8510.2420.7300.2900.523

 Intermittent service (piped supply only)

 23

 2 

 0.52 (0.12 - 2.3)

 0.390

Functional problems, well 6 4 

4.5 (1.2 - 17) 0.024

 Free chlorine between 0.2 and 2 mg/l

 146

 7

 0.16 (0.069 - 0.37)

 <0.001

Mean free chlorine (mg/l) 0.691 0.164    

WELL WATER WITH SIGNIFICANTLY HIGHER FC

WELL PROBLEMS ASSOCIATED WITH FC

Cl RESIDUALS ASSOCIATED WITH LESS FC

   FC <1 cfu/100ml 

 FC ≥ 1 cfu/100ml

 OR (95% CI)

 p-value

Number of people 450 57    

Mean age 34.7 37.1    

Self-reported symptoms, 7 day

       

Diarrhea 6 4 5.6 (1.5 - 20) 0.009

Abdominal cramps 11 2 1.5 (0.31 - 6.7) 0.634

Nausea 10 3 2.4 (0.65 - 9.2) 0.185

Vomiting 7 1 1.1 (0.14 - 9.4) 0.910

Fever 5 4 6.7 (1.8 - 26) 0.006

 HCGI*

 8

 4

 4.0 (1.2 - 14)

 0.027

*73 people reported pre-existing conditions – diverticulitis, heartburn, irritable bowel syndrome, milk intolerance, stomach ulcer, colitis, or migraine - these are excluded

   FC <1 cfu/100ml 

 FC ≥ 1 cfu/100ml

 OR (95% CI)

 p-value

Number of people 450 57    

Mean age 34.7 37.1    

Self-reported symptoms, 7 day

       

Diarrhea 6 4 5.6 (1.5 - 20) 0.009

Abdominal cramps 11 2 1.5 (0.31 - 6.7) 0.634

Nausea 10 3 2.4 (0.65 - 9.2) 0.185

Vomiting 7 1 1.1 (0.14 - 9.4) 0.910

Fever 5 4 6.7 (1.8 - 26) 0.006

 HCGI*

 8

 4

 4.0 (1.2 - 14)

 0.027

   FC <1 cfu/100ml 

 FC ≥ 1 cfu/100ml

 OR (95% CI)

 p-value

Number of people 450 57    

Mean age 34.7 37.1    

Self-reported symptoms, 7 day

       

Diarrhea 6 4 5.6 (1.5 - 20) 0.009

Abdominal cramps 11 2 1.5 (0.31 - 6.7) 0.634

Nausea 10 3 2.4 (0.65 - 9.2) 0.185

Vomiting 7 1 1.1 (0.14 - 9.4) 0.910

Fever 5 4 6.7 (1.8 - 26) 0.006

 HCGI*

 8

 4

 4.0 (1.2 - 14)

 0.027

ODDS OF HCGI ARE 4x HIGHER IF WATER SAMPLE IS POSITIVE FOR FECAL COLIFORM

Summary: Alabama, Phase 1

• Systems and wells at risk for microbiological contamination

• Onsite sanitation a (potentially big) problem• High prevalence of gastro-intestinal illness

– Suggestive evidence that water may play a role

• Need for assistance with DBP rules

Phase 2: Current study• Water Infrastructure

Sustainability and Health in Alabama's Black Belt – Advancing Public Health

Protection through Water Infrastructure Sustainability (EPA-G2009-STAR-F1)

• 2011 – 2015, 14 water systems, 3 counties

Goals: Phase 2• 1. Assess system performance and

water quality– Fecal indicators and pathogens

• 2. GI illness associations – Directly measured (epidemiology)– Modeled using QMRA

• 3. Identify possible transmission pathways

• 4. Identify risk mitigation strategies– Working with operators

Methods• Household-level water quality data

– Microbiological indicators, pressure, Cl, pH, turbidity

• Health GI measures– HCGI health diaries & interviews– Salivary antibody assay pilot

• Pathogens and indicators from concentrated ultrafiltration samples– Cryptosporidium, Giardia, E. histolytica,

norovirus, rotavirus, adenovirus, Enterococcus, E. coli, Campy, Salmonella

• Other system-level data– System modeling – Operation and maintenance– Infrastructure characteristics

Methods, continued• Survey of 900 randomly selected

connected households (300 in each county)– Health outcomes– Grab vs flamed drinking water samples – Septic access and function– Perceptions and water use

• Ongoing surveillance• Selection of sampling sites

– Selected households, key system points

– Up to 10 sample points per system• Dead-end ultrafiltration

Turbidity – cross-sectional TC– cross-sectional

Selecting sampling points

System data• Flow modeling • O&M records• Existing monitoring data• Offering PDHs to

operators

Onsite sanitation• Surveys of OSS function and access

– Public records vs survey– Previous study: 90% of septic

systems in the Black Belt were failing (ADPH 1997)

• Microbial source tracking of E. coli isolates– Household-level samples, drainfield

samples, system samples– Partner study: household well

samples• Technology innovation for

underserved communities– Urgent need

COURTESY OFR GELTING + CDC