Drowning Interventions, Pool Fencing

Background

Over the last decade, many communities have passed laws requiring the fencing of private and public pools. In fact, much of the evidence that pool fencing reduces the risk of drowning among children comes from studies examining before and after rates of drowning for fenced and unfenced pools. Even after these pioneering studies had been done, the favorable attitude toward pool fencing did not necessarily translate into actual changes in pool fencing.6

More recently, studies have examined whether the type of fence surrounding a pool makes a difference. Comparison studies7 of perimeter fencing (property barrier) versus isolation fencing (around immediate pool area) show isolation fencing to be much more effective in reducing the risk of drowning. Adding to this finding are studies8,9 examining children’s ability to climb certain types of fences. In particular, chain-link fences, while allowing visibility of the pool area, are the most easily scaled by children as young as 2 years old. Ornamental iron bar fences were advocated as a better barrier, with reduced ‘climability’ while retaining the visibility factor.

Fence height makes little difference if the child is able to climb it; one study8 showed the median time for four-year-olds to climb a 5-foot fence was seventeen seconds. For a fence that is not scaleable, the most important element of fencing is a secure, self-closing gate. 

Taking these and other findings into account, the U.S. Consumer Product Safety Commission has compiled a list of minimum recommendations10 for residential pool fencing; recent legislation passed in Seattle11 requires a fence height of at least 5 feet, and an inter-bar spacing of no more than 4 inches.

Here we review the relevant studies which have examined the effect of pool fencing on drowning. These studies are time series, case-control, or non-equivalent control group comparisons. No randomized controlled studies have been conducted.


Review of pool fencing intervention studies:
 

Author

Morgenstern et al., 2000

Study design and target population

Retrospective cohort study of pool drownings in Los Angeles County, CA.

(Stage 1), case control study.

(Stage 2), cases-all pools where drownings occurred, 1990-95; controls-randomly selected pools matched by geographic & ecological areas.

Intervention

Presence of pool fencing ordinance when pool built or upgraded.

Outcomes

Residential pool drownings in children under 10 years of age.

Results

146 deaths in 6 years.

1.77 per 100,000 children per year; rate 9-10 times higher in 1-4 year old children than in infants or school age kids.

81% of all drownings occurred in pools that were regulated by pool-fencing ordinances.

OR=1.27 (95% CI=0.72, 2.25) comparing presence of fencing ordinance to absence of fencing ordinance.

Study quality and conclusions

Implies that laws governing fencing are ineffective.

Study couldn't distinguish between above-ground and in-ground pools.

Local ordinances did not require 4-sided fencing.

Inadequate enforcement of ordinance and poor maintenance of pool barriers.

Future studies should include exposure data (person-time at risk).



Author

Pitt & Balanda, 1998

Study design and target population

Time series; Queensland Australia.

Intervention

Uniform pool fence requirements, 1992.

Outcomes

Drowning in residential pools, children 1-4 years of age.

Results

Deaths decreased from 12/year (1982-1991) to 2/year (1992-1994) but increased 11/year (1995-1997).

Increased exposure, 70,000 additional pools added between 1991 and 1997.

33 of 38 pools (87%) where children drowned did not comply with fencing legislation.

Study quality and conclusions

Good surveillance databases & investigation of cases.

Recommend policing to improve compliance with legislation and public education campaign.



Author

Pitt et al., 1991

Study design and target population

Population-based case-control study

All children 0-13 years old with immersion injury in Brisbane (n=139) from 1984 to 1989.

Cases: Children with immersion injury from unintentional access to domestic pools

Controls: Random sample of households with pools

Intervention

Pool fencing stratified by type (isolation, three-sided, none)

Outcomes

Comparison of drowning and near-drowning rates for fenced and unfenced pools.

(Because all children gaining unintentional access to pool from three-sided fencing did so through the house door, both no fencing and three-sided fencing comprise the ‘unfenced’ category.)

Results

Significantly decreased risk of drowning in isolation-fenced pool compared to unfenced pool (OR=0.27, 0.15-0.49).

Similar, significantly decreased risks when stratified by pool type (inground, above-ground, spa).

Eighty-three percent of all immersions occurred in children less than 3 years of age.

Study quality and conclusions

All analyses restricted to immersions with unintentional access to pool.


Author

Intergovern-mental Working Party on swimming pool safety (Australia), 1988

Study design and target population

Population based case-control study

Preschoolers aged 0-4 years in Metropolitan Perth, Australia, 1975-Feb. 1988.

Cases: Deaths from immersion

Controls: Pool fencing type among households with children 0-4 years

Intervention

Perimeter fencing v. isolation fencing.

Outcomes

Incidence rates for drowning for specific fencing types; AR%s.

Swimming Pool Drowning Index (SPDI), adjusting for population distribution and home pool installation changes (i.e., exposure to pools) over time. 

Results

Significantly decreased risk of drowning in isolation-fenced pool compared to perimeter-fenced pool (OR=0.16, 0.05-0.49); AR%=0.84.

Study quality and conclusions

Good ascertainment of drownings.

SPDI for near-drownings more important from public health view, but difficult to ascertain number of private pools outside city limits.



Author

Present, 1987

Study design and target population

Case-control study

Reported drownings and near-drownings to children under age five in eight selected U.S. counties

Cases: Children with immersion injury in domestic pools

Controls: Non-random population survey of households with pools

Intervention

Isolation fencing v. three-sided fencing (access from house or no fencing).

Outcomes

Comparison of drowning cases by fencing type

Results

Logistic regression showed a non-significant, negative relationship between isolation fencing and drowning (data not presented).

Study quality and conclusions

Controls matched on household income and type, and age of proxy interviewee. Controls self-selected. Incomplete case ascertainment. 


Author

Nixon et al., 1986

Study design and target population

Interrupted time series study

Reported drownings of children in Brisbane from 1967 to 1981. All near-drownings from 1971-1981.

Intervention

Broad assessment of educational and public awareness programs (daily newspaper articles, letters to the editor, news reports).

Brief mention of legislative safety intervention requiring pool fencing (type not specified).

Outcomes

Comparison of drowning and near-drowning annual incidences.

Results

Safety legislation appears to decrease drowning incidence by about one-third.

Media campaigns may affect whole drowning spectrum.

Actual decrease in survival rate after safety campaign.

Study quality and conclusions

Type of safety legislation not stated. Combined drowning and near-drowning rates did not show as much decline as drowning rates did. No adjustments made for age, population distribution changes over time, or number of pools.


Author

Fergusson et al., 1984

Study design and target population

Case-control study

Drownings in fenced and unfenced pools estimated from available New Zealand-based data

Cases: Domestic drownings where fencing type was known

Controls: Estimates of domestic pool fencing types

Intervention

Pool fencing versus no fencing (type not specified).

Outcomes

Risk of drowning in unfenced pool and attributable risk percent calculated using Bayes’ Theorem.

Results

Risk of drowning in fenced pool significantly lower than for unfenced pool (OR range=0.21–0.48)

Reduction in risk, expressed as attributable risk percent, ranges from 41% to 67%.

Study quality and conclusions

"Cases" derived from probabilities cited in previous studies and personal communication.

Quality of previous studies unknown. Relatively small numbers may make risk estimates unstable.

No definition of fenced/unfenced pools.



Author

Milliner et al., 1980

Study design and target population

Ecological (case series) study

Children 0-14 years old.

Intervention

Legislation requiring pool fencing (apparently isolation fencing, defined as barriers between pool and house, but unclear in article); enforcement of pool-fencing law.

Outcomes

Comparison of drowning, near-drowning requiring admission or ED care in Mulgrave (with law) to Cairns (without law).

Results

Fatality rate using mid-survey population is 7.7/100,000 in Mulgrave (with law), compared to 14.3/100,000 in Cairns (without law). RR=0.54

Near-drowning rate is 15.4/100,000 in Mulgrave, compared to 14.3/100,000. RR=1.08

Study quality and conclusions

Population estimates are for all ages; drownings and near-drownings for children 0-15 years old. Near-drowning rates include rivers.

All 10 drownings in Cairns Shire occurred in unfenced pools. One drowning in unfenced pool (previously exempted) in Mulgrave Shire.



Author

Pearn et al., 1979

Study design and target population

Ecological comparison study (no before/after comparison of legislation)

Children and adolescents under 16 years old.

Intervention

Regulation requiring swimming pool fencing (type not specified).

Outcomes

Swimming pool fatality rate.

Results

Fatality rate is 0.9/100,000 in Honolulu (with law) compared to 2.6/100,000 in Brisbane (without law). RR=0.35

Study quality and conclusions

No controlling for exposure. 

Odds ratio summary table (image)

Summary of pool fencing studies

Early studies of pool fencing compare drowning rates in communities with and without fencing legislation, revealing significantly higher drowning incidences in those areas without any pool fencing laws. However, this evidence is gathered from ecological studies or case series, rendering the preventive association between fencing and drowning tentative at best. Subsequent case-control studies are able to examine pool fencing intervention (within and between communities), showing that pool fencing significantly reduces the risk of drowning. Case control studies evaluating fencing type indicate that isolation fencing (enclosing pool only) is superior to perimeter fencing (enclosing property and pool) because perimeter fencing allows access to the pool area through the house. The two recent studies evaluating pool fencing ordinances in Los Angeles County and Queensland, Australia indicate that the mere passage of legislation is not sufficient to reduce drownings.

To reference the complete Cochrane Review on Pool Fencing click here.

Recommendations on effectiveness of pool fencing and regulations requiring pool fencing

There is no reason to believe that a particular fencing type shown to be effective in preventing toddlers from climbing it should not translate to the real world. Because three-sided pool fencing still allows for unintentional access through the house, we therefore strongly recommend isolation pool fencing as an effective strategy in preventing submersion injuries. Legislation requiring fencing does increase the percentage of pools that are fenced. However, many children and toddlers gain access to pools from a poorly-latched gate or from the house itself if the fence is three-sided. Pool fencing that serves merely as a barrier to trespassers will not prevent drownings for those allowed access to the property. Pool fences should have a dynamic and secure gate and isolate (i.e., four-sided fencing) the pool from the house. Most unintentional accesses to private pools with isolation fences occur because of a poorly-working or poorly-designed latch, or because a toddler follows older children inside the pool fencing. Legislation should require isolation fencing with secure, self-latching gates. Legislation should require retrofitting of existing pools and include enforcement provisions in order to be effective.

Recommendations for future research

Studies examining fencing enforcement might lead to better fencing legislation. Specifically, case-control study comparing types of fencing legislation and their policies (fines, periodic inspections, etc.) would be important in determining the most effective and practical means of enforcement for a given community. Future studies should examine the effect of adding a multifaceted educational campaign to legislation. Creating, updating, and maintaining surveillance databases of drownings are essential in order to evaluate the effectiveness of fencing type or the effectiveness of any other intervention. These databses should include near-drownings, information on circumstances of drowning, number of private and public pools, person-time exposure measures, fencing types, and regulations.