The connection between housing and health is well established.1–6 Homes provide shelter and security, but homes that are poorly constructed or maintained can have a significant impact on the health and safety of residents. For example, asthma continues to account for millions of hospital and emergency department visits every year, carbon monoxide (CO) exposure is a leading cause of poisoning-related deaths in the United States, and, despite decades of investment, an estimated half a million children between 1 and 5 years of age have elevated blood lead levels.7–10 These issues and many others are directly linked to poor housing conditions, which are underscored by the fact that geographic, income, and racial disparities in housing quality mirror those of some of our most significant public health challenges.11–15 These are also costly problems for our health care system and society, accounting for billions of dollars in health care expenditures and lost productivity every year.16
Over the past decade, multiple agencies have worked to advocate for an increase in services, programs, and policies that promote healthy housing.4,17–20 These calls to action have cited evidence from research and demonstration projects that highlight the potential public health impact of improved housing quality. Traditionally, many housing hazards have been addressed through initiatives targeted at specific health outcomes or environmental exposures (eg, remediation of lead hazards), but there has been a growing shift toward a more comprehensive approach that addresses multiple housing deficiencies within a single program.1,3,4,17,20–24 This approach is premised on the experience that homes with one type of hazard often have others, and remediation of an underlying problem can impact more than 1 endpoint. For example, correction of a moisture problem may not only prevent deterioration of lead-based paint but will also aid in preventing mold, pests, and structural disrepair. Furthermore, since gaining entry to a home to complete a health and safety assessment can be very resource-intensive, proponents argue that it is more cost-effective to address multiple issues within a single program.
However, as the evidence base has grown, some critical questions remain about the feasibility and promise of implementing healthy homes approaches at scale and in real-world settings. What is the impact of improved housing on different populations (eg, children vs adults, urban vs rural)? Which hazards can be improved? How much improvement can we expect from a healthy homes program operating to scale? Can we expect improvements from low-intensity interventions that do not provide structural improvements? Will modest improvements in housing conditions yield improvements in health that have clinical or public health significance? Are these types of programs sustainable? While existing reports from research-based designs, demonstration projects, and smaller-scale initiatives have contributed to our current understanding of the impact of healthy homes approaches,3,13,22–28 less is known about the impact of large-scale programs operating in real-world settings.
The New York State (NYS) Healthy Neighborhoods Program (HNP) has operated since 1985 and is distinct from previously evaluated programs in its scale, geographic scope, and funding structure. In this report, we present findings from a retrospective evaluation of this large-scale, multisite, state-funded healthy homes initiative. Although the program is nearly 30 years old, the evaluation focuses on a recent 5-year period after the implementation of a standardized data collection instrument. This unique data set includes information about services provided to more than 29 000 homes with 82 000 residents in 13 counties. We report on the proportion of homes affected by each type of housing hazard, improvement in hazards among revisited homes, and the change in the overall number of hazards per home.
A select number of county health departments are funded by the NYS Department of Health (DOH) to provide in-home assessments and interventions to improve the environmental health and safety of residents. During the evaluation time frame (2008-2012), 13 counties were funded from the state's general funds, 10 continuously, to implement this program. During this time frame, grants totaled $10 million, about $345 per visited dwelling. The counties are spread throughout state and vary in urbanicity, ranging in population density from 79 persons per square mile along the border with Canada to 2205 persons per square mile just north of New York City.
Each county selects high-risk areas to target, including specific zip codes in urban areas, towns, or geographic regions. Homes and residents are approached through a combination of door-to-door canvassing and referrals from other programs, local organizations, or health care providers. Trained environmental health specialists (sanitarians, health educators, public health nurses, or other public health professionals) provide the intervention, which varies in scope according to the needs identified during the assessment and the specific expertise of the staff. All homes receive an initial visit, which includes a visual assessment to identify potential health and safety hazards and the provision of guidance, products, and referrals to address any identified hazards. The assessment and interventions address tobacco control, fire safety, lead poisoning prevention, indoor air quality (CO, radon, ventilation, odors, temperature, and humidity), general conditions (cleaning, clutter, pests, mold/mildew, moisture, and structural problems), asthma triggers, and others (eg, injury prevention and social services). Three to 6 months after the initial visit, the counties are expected to conduct revisits for roughly a quarter of all homes. During a revisit, the home is reassessed and any new or ongoing problems are addressed. The timing and selection of households for revisit are determined by each county program, but preference is typically given to homes with pressing health or safety issues or residents with asthma.
The HNP is designed to provide a framework of core objectives, operating procedures, and measures that are consistent across all programs. However, the counties are encouraged to build on local resources and infrastructure to deliver services in a way that is meaningful and effective for each community. For example, pest control problems may be referred to the county's vector control program or to a local university's extension service.
In terms of remediation intensity, the HNP is best classified as a minor to moderate intervention.19 While program participants are provided with advice, low-cost materials (eg, pest control baits, mattress covers), and referrals for services, the program does not provide or fund structural improvements, repairs, or professional cleaning services and is therefore considered a lower-intensity approach. By addressing multiple housing hazards in 1 visit, this approach meets the definition of a healthy homes approach that is comprehensive or holistic (in contrast to programs that address a single exposure or disease).29
The program evaluation of the HNP was reviewed by the institutional review board of the NYS DOH and given exempt status on the basis of not being research.
Assessment instrument and data
The data used in this analysis were collected on a standardized home assessment form (see Document, Supplemental Digital Content 1, available at: http://links.lww.com/JPHMP/A285). The form includes 42 hazards that are assessed during the initial visit and revisit (see Document, Supplemental Digital Content 2, available at: http://links.lww.com/JPHMP/A286). Staff are provided with a manual that includes guidelines and strategies for assessing the presence of a hazard and instructions for completing the assessment form (see Document, Supplemental Digital Content 3, available at: http://links.lww.com/JPHMP/A287). The assessment covers 5 basic areas of home health and safety: tobacco control, fire safety, lead, indoor air quality, and other general conditions. The form is completed by the surveyor during or immediately after each home visit. The main form includes demographic information about the primary respondent, characteristics of the dwelling, enumeration and characteristics of the residents, physical conditions of the dwelling, and education, referrals, and products that were provided. An asthma form is completed at each visit for each person reporting a doctor diagnosis of asthma and includes information about the presence of asthma triggers in the home, asthma symptoms and morbidity, and asthma self-management. The completed forms are faxed to NYS DOH, and the faxed image is scanned and saved to a database (without personal identifiers). The data fields are automatically checked for completeness and valid values, and errors are manually verified and corrected.
To describe the prevalence and range of hazards succinctly, we set out to develop a score that could, in a single number, characterize the extent of the health and safety hazards in a home. We convened an expert panel of statisticians and health and housing experts and used a modified nominal group technique to select hazards for inclusion in the score. All 42 hazards assessed were considered for inclusion in the home hazard score. Participants were asked to consider each hazard on 3 sets of factors: nexus to health (including strength of association with health outcomes and potential severity of outcomes), likelihood of an adverse event, and the relative ease or difficulty of improving the hazard and maintaining the improvement. Individuals independently assigned hazards for inclusion, possible inclusion, or exclusion from the score, followed by discussion to make the final selection of that would comprise the score, possible combinations of related individual hazards, and how to handle missing observations. The home hazard score includes 19 elements: 15 individual hazards and 4 combinations of related hazards. For example, “rodents” is the presence or evidence of mice and/or rats. After considering a weighting scheme, the expert panel decided that all 19 hazards would be weighted equally but decided that further weighting was not warranted because no information about the extent or magnitude of the hazard was collected.
We selected dwellings that had an initial visit in 2008-2012 where at least 15 of the 19 hazards had been assessed. To be included in the analysis of the change from the initial visit to the revisit, the dwellings had to have a revisit 60 to 240 days after the initial visit and have at least 15 hazards assessed at the revisit. The final sample included 28 491 dwellings with initial visits and 6436 dwellings with both an initial visit and a revisit.
Home hazard score
Hazards assessed during the initial visit are recorded as present or absent; during the revisit, hazards are recorded as present, absent, or improved. For this analysis, if a hazard was present at the initial visit and absent or improved at the revisit, the hazard was classified as improved. Each hazard present in the home was assigned a weight of 1, and each that was absent was assigned a weight of zero. For homes with 4 or fewer missing hazards, the prevalence of the hazard was used as follows. If a hazard was missing at the initial visit, the weight was set to the initial visit prevalence of the hazard among the nonmissing observations. If a hazard was missing at the revisit but was reported at the initial visit, the weight was set to the conditional prevalence that the hazard was present at the revisit, given the observation at the initial visit. If a hazard was missing at both the initial visit and the revisit, the revisit weight was set to the revisit prevalence of the hazard among the nonmissing observations.
The home hazard score is the sum of the weights of the hazards at each visit. Subtracting the revisit score from the initial visit score is the change; a negative difference indicates a reduction in the score or improvement in the conditions assessed. We also calculated the sum of the number of residents living in a home with 1 or more hazards at the initial visit and estimated the number of residents impacted by improvements by assuming that the proportion of homes where 1 or more hazards were corrected was the same whether or not a home was revisited.
We used McNemar's test to assess whether the percentage of homes with the individual hazard changed from the initial visit to the revisit. For each dwelling, we categorized the change in the scores as improved, worsened, and no change. We used a weighted least squares test to assess whether the percent improved was different from the percent worsened. To test that there was a change in the mean scores from the initial visit to the revisit, we used a paired t test. The signed-rank test was used to test that median change in scores was nonzero. Statistical significance for all tests was defined as P < .05. All analyses were conducted with SAS (version 9.4; SAS Institute, Inc, Cary, North Carolina).
The demographics of the dwellings, primary respondents, and households are presented in Table 1. Of the dwellings with an initial visit (N = 28 491), most had 1 to 5 units, two-thirds were built before 1950, and two-thirds were rental units. About 40% of primary respondents were black or African American, 16% were Hispanic, and about three-fourths had a high school diploma. Half of the households received public assistance. The percentages were similar for the dwellings with a revisit (N = 6436).
Table 2 lists the prevalence of the home environmental hazards that were included in the home hazard score. Although the prevalence of hazards at the initial visit differed between all dwellings with an initial visit and the dwellings that also had a revisit, most were within 2 percentage points. In the revisited dwellings, the most prevalent conditions at the initial visits were a missing CO detector (68%), missing smoke detectors (40%), and smoking in the home (39%), followed by odors of chemicals or scented products (24%), ineffective cleaning/dust accumulation (17%), rodents (14%), and significant clutter (13%). Twenty-seven percent of homes were built before 1978 and had a lead paint hazard inside or outside the home.
Among conditions included in the home hazard score, all but chemical smells were significantly improved after the interventions (P < .05). The highest percentages of improved homes were for missing smoke detectors (95%), missing CO detectors (76%), malfunctioning appliances (68%), blocked exits (67%), rodents (65%), cockroaches (58%), leaks (54%), electrical hazards (54%), and mold (53%).
Table 3 presents the summary statistics for the home hazard score. The mean home hazard score at the initial visit for all dwellings was comparable with the dwellings with a revisit (2.7 and 2.8, respectively). In addition, 9% of all dwellings had a home hazard score of zero (ie, 91% had a nonzero score) compared with 7% of the revisited dwellings. The mean change in the home hazard score was −1.3 (a reduction of 1.3 hazards per home, P < .001). The home hazard score decreased in 75% of homes and increased in 5% (P < .001). A total of 74 479 residents were living in homes affected by 1 or more types of housing hazards at the initial visit. Of these, we estimate that 59 483 (80%) were impacted by improvements in housing conditions after the intervention.
Within a short follow-up period, the NYS HNP was able to reduce the overall number of hazards per home and demonstrate significant improvements in the conditions assessed for fire safety, indoor air quality, tobacco control, lead poisoning prevention, pest control, mold and moisture, and other environmental health and safety hazards in the homes of residents in high-risk neighborhoods.
Fire safety, lack of a CO detector, and smoking in the home were the most common types of hazards identified. With the exception of chemical smells, all 29 hazards, including all 19 used in the hazard score, were significantly improved following the intervention. Hazards associated with fire safety, CO safety, pests, leaks, and mold showed the greatest magnitude of improvement. Homes generally had more than 1 environmental health and safety hazard identified at the initial visit, but, despite a reduction in the overall number of hazards, were not hazard-free at the revisit.
Others have previously reported on similar efforts to assess and reduce environmental health and safety hazards in the home environment. These programs vary in the scope of what is addressed, the size and characteristics of the populations served, and the intensity of the interventions provided, which can make direct comparisons challenging. The proportion of homes impacted by specific hazards at the initial visit varies greatly across published studies, perhaps reflecting differences both in geography and in the population recruited.21,23,30–34 However, reports by Dixon et al21 and Klitzman et al23 affirm our findings that homes typically have more than 1 hazard and are not hazard-free at the revisit but reported a greater number of hazards at the initial visit.
In general, the magnitude of improvement reported by the NYS HNP was either comparable with or more modest than improvements reported by other programs, but direct comparisons are again complicated by the differences in prevalence at the initial visit and in the intensity of the intervention offered.21,23,35,36 However, in contrast to the HNP, previous studies with smaller sample sizes, including some that reported higher magnitudes of improvement for individual outcomes, often failed to demonstrate that their outcomes were statistically significant. In terms of the home hazard score, both Dixon et al21 and Klitzman et al23 reported more dramatic improvements in the overall number of hazards per home than our study, perhaps reflecting differences in the types of hazards counted and in the intensity and focus of the intervention provided.
The reduction in the overall number of hazards per home provides evidence in support of a comprehensive healthy homes approach that addresses multiple hazards in a single visit. In developing the home hazard score, we also examined, but did not present, the bivariate relationships between all pairs of the 29 hazards in Table 2 (see Table, Supplemental Digital Content 4, available at: http://links.lww.com/JPHMP/A288). Of 406 associations tested, 331 were significantly positively associated (with odds ratios ranging from 1.2 to 108.6), 22 were significantly negatively associated (with odds ratios ranging from 0.3 to 0.5), and 41 did not have a significant association. In keeping with the healthy homes premise, positive associations confirmed that not only one type of hazard is often significantly associated with a range of other possible outcomes (eg, mold, pests, and lead paint hazards were all significantly associated with leaks) but also those outcomes are associated with each other (eg, lead paint hazards were positively associated with mold). Of note among the negative associations were the relationships between odors from scented products and ineffective housecleaning, dust accumulation, clutter, and rodents. Although odors from scented products are often considered an asthma trigger, people who use them may maintain cleaner homes.
Strengths and limitations
This evaluation is important and unique, particularly due to its large sample size, geographically diverse population, and real-world setting. However, there are also a number of important limitations of this evaluation. The short follow-up period prevents us from assessing longer-term sustainability of improvements in housing conditions, and the lack of a comparison group precludes us from attributing the entirety of the observed improvements to the intervention. The flexible protocol means there may be important differences in the way that homes are assessed and problems addressed, and it was not possible to attribute improvements to individual intervention components (eg, education vs referrals). While this flexibility is advantageous for providing services to residents in a variety of settings, the range of challenges faced by residents in HNP communities (in housing conditions, occupant behaviors/attitudes, resources available) may not be representative of communities elsewhere. Conversely, aggregating findings across counties may mask geographic variation in housing conditions and improvements.
The development of the hazard score was constrained by the list of hazards assessed by the program. Other efforts to develop similar scores have also relied on a combination of evidence from the peer-reviewed literature and expert opinion, but, in contrast to our approach, the ability to generate a hazard score influenced the design of the assessment.21,23,25,34,37,38
Finally, the evaluation is subject to bias, including selection bias (which homes allow access, which are targeted for revisit), recall bias (resident reporting of childhood lead screening), social desirability bias (for issues such as tobacco use or presence of cockroaches), and reporting bias (surveyors' subjectivity in evaluating their own work at the revisit). However, no differences were observed between the improvements reported when a different surveyor conducted the revisit (data not shown) and the impact of social desirability bias is mitigated by a reliance on visual assessment (eg, the presence of ashtrays outweighs a statement that no smoking occurs in the home). While some conditions are more easily assessed than others, evidence suggests that visual assessment, although not as rigorous as environmental sampling, is an acceptable proxy for estimating the potential presence of allergens and irritants in the home environment.39
Implications for Policy & Practice
* As resources for public health interventions are increasingly constrained, lessons learned from healthy homes programs such as the NYS HNP may be of great value to decision makers.
* One way to increase provision of healthy homes services is to expand the focus of existing programs so that they address multiple housing issues.
* This evaluation provides information about hazards that are common, easily assessed, and easily corrected or improved, which may be of use to a variety of programs that already provide in-home services and are seeking to expand the scope of their visits.
* In addition, these findings can inform the development of new healthy homes programs and policy so that both resources and public health benefits are maximized.
This evaluation of the NYS HNP adds to the evidence that a low-intensity healthy homes approach can be used to achieve striking improvements in housing conditions that have public health impact for both adults and children and across diverse geographic settings. Future evaluations of this and similar programs should attempt to assess long-term sustainability, associations with a wider range of health outcomes, attribution of improvement to intervention components (eg, health education, use of provided products or referrals), comparative impact of higher-intensity interventions, and cost-effectiveness of the full range of home environmental improvements.
1. Krieger J, Higgins DL. Housing and health: time again for public health action. Am J Public Health. 2002;92(5):758–768.
2. Lowry S. Housing. BMJ. 1991;303:838–840.
3. Matte TD, Jacobs DE. Housing and health—current issues and implications for research and programs. J Urban Health. 2000;77(1):7–25.
4. National Center for Healthy Housing. Housing Interventions and Health: A Review of the Evidence. Columbia, MD: National Center for Healthy Housing; 2009. http://http://www.nchh.org
/LinkClick.aspx?fileticket=2lvaEDNBIdU%3d&tabid=229. Accessed January 28, 2016.
5. Sandel M, Phelan K, Wright R, Hynes H, Lanphear B. The effects of housing interventions on child health. Pediatr Ann. 2004;33:474–481.
6. Jacobs DE, Wilson J, Dixon S, Smith J, Evens A. The relationship of housing and population health: a 30-year retrospective analysis. Environ Health Perspect. 2009;117:597–604.
7. Centers for Disease Control and Prevention. Nonfatal, unintentional, non-fire-related carbon monoxide exposures—United States, 2004-2006. MMWR Morb Mortal Wkly Rep. 2008;57:896–899.
8. Landrigan PJ, Schecter C, Lipton J, Fahs M, Schwartz. Environmental pollutants and disease in American children: estimates of morbidity, mortality, and costs for lead poisoning, asthma, cancer, and developmental disabilities. Environ Health Perspect. 2002;110:721–728.
9. National Center for Environmental Health. Lead. http://http://www.cdc.gov
/nceh/lead. Published 2011. Accessed 2011.
10. National Heart, Lung, and Blood Institute. Morbidity and mortality chartbook. http://http://www.nhlbi.nih.gov
/resources/docs/cht-book.htm. Published 2009. Accessed 2011.
11. Bashir SA. Home is where the harm is: inadequate housing as a public health crisis. Am J Public Health. 2002;92:733–738.
12. Hood E. Dwelling disparities: how poor housing leads to poor health. Environ Health Perspect. 2005;113:A310–A317.
13. Miller WD, Pollack CE, Williams DR. Healthy homes and communities: putting the pieces together. Am J Prev Med. 2011;40:S48–S57.
14. Rauh VA, Landrigan PJ, Claudio L. Housing and health: intersection of poverty and environmental exposures. Ann N Y Acad Sci. 2008;1136:276–288.
15. Raymond J, Wheeler W, Brown MJ; Centers for Disease Control and Prevention. Inadequate and unhealthy housing, 2007 and 2009. MMWR Suppl. 2011;60(1):21–27.
16. Trasande L, Liu Y. Reducing the staggering costs Of environmental disease in children, estimated at $76.6 billion in 2008. Health Aff. 2011;30(5):863–870.
17. Brown MJ, Ammon M, Grevatt P. Federal agency support for healthy homes. J Public Health Manag Pract. 2010;16(5)(suppl):S90–S93.
18. US Department of Health and Human Services. The Surgeon General's call to action to promote healthy homes. http://www.ncbi.nlm.nih.gov
/bookshelf/br.fcgi?book=ctahome. Published 2009. Accessed 2010.
19. Centers for Disease Control and Prevention. Guide to Community Preventive Services: Asthma Control. http://www.thecommunityguide.org
/asthma/index.html. Published 2008. Accessed 2011.
20. US Department of Housing and Urban Development. Leading our nation to healthier homes: the healthy homes strategic plan. http://http://www.hud.gov
/offices/lead/library/hhi/hh_strategic_plan.pdf. Published 2009. Accessed 2016.
21. Dixon SL, Fowler C, Harris J, et al An examination of interventions to reduce respiratory health and injury hazards in homes of low-income families. Environ Res. 2009;109:123–130.
22. Jacobs DE, Brown MJ, Baeder A, et al A systematic review of housing interventions and health: introduction, methods, and summary findings. J Public Health Manag Pract. 2010;16(5)(suppl):S5–S10.
23. Klitzman S, Caravanos J, Belanoff C, Rothenberg L. A multihazard, multistrategy approach to home remediation: results of a pilot study. Environ Res. 2005;99:294–306.
24. Saegert SC, Klitzman S, Freudenberg N, Cooperman-Mroczek J, Nassar S. Healthy housing: a structured review of published evaluations of US interventions to improve health by modifying housing in the United States, 1990-2001. Am J Public Health. 2003;93:1471–1477.
25. Breysse P, Farr N, Galke W, Lanphear B, Morley R, Bergofsky L. The relationship between housing and health: children at risk. Environ Health Perspect. 2004;112:1583–1588.
26. Brugge D, Farr N, Galke W, Lanphear B, Morley R, Bergofsky L. How much evidence is enough? Assessing home asthma research. J Public Health Manag Pract. 2010;16(5)(suppl):S21–S23.
27. Jacobs DE, Kelly T, Sobolewski J. Linking public health, housing, and indoor environmental policy: successes and challenges at local and federal agencies in the United States. Environ Health Perspect. 2007;115:976–982.
28. Sandel M, Baeder A, Bradman A, et al Housing interventions and control of health-related chemical agents: a review of the evidence. J Public Health Manag Pract. 2010;16(5)(suppl):S24–S33.
29. National Center for Environmental Health. CDC's Healthy Homes Initiative. http://http://www.cdc.gov
/nceh/lead/healthyhomes.htm. Published 2009. Accessed 2011.
30. Centers for Disease Control and Prevention. 2006 Surgeon General's report—the health consequences of involuntary exposure to tobacco smoke. http://www.cdc.gov
/tobacco/data_statistics/sgr/2006/index.htm. Published 2006. Accessed 2010.
31. Jacobs DE, Clickner RP, Zhou JY, et al The prevalence of lead-based paint hazards in U.S. housing. Environ Health Perspect. 2002;110:A599–A606.
32. Nguyen T, Lurie M, Gomez M, Reddy A, Pandya K, Medvesky M. The National Asthma Survey—New York State: association of the home environment with current asthma status. Public Health Rep. 2010;125:877–887.
33. Runyan CW, Johnson R, Yang J, et al Risk and protective factors for fires, burns, and carbon monoxide poisoning in U.S. households. Am J Prev Med. 2005;28:102–108.
34. Takaro T, Krieger J, Song L, Sharify D, Beaudet N. The Breathe-Easy Home: the impact of asthma-friendly home construction on clinical outcomes and trigger exposure. Am J Public Health. 2011;101(1):55–62.
35. Schwartz L, Martinez L, Louie J, et al An evaluation of a carbon monoxide poisoning education program. Health Promot Pract. 2010;11:320–324.
36. Ta VM, Frattaroli S, Bergen G, Gielen A. Evaluated community fire safety interventions in the United States: a review of current literature. J Community Health. 2006;31:176–197.
37. Krieger JW, Takaro TK, Song L, Weaver M. The Seattle-King County Healthy Homes Project: a randomized, controlled trial of a community health worker intervention to decrease exposure to indoor asthma triggers. Am J Public Health. 2005;95(4):652–659.
38. Nriagu J, Smith P, Socier D. A rating scale for housing-based health hazards. Sci Total Environ. 2011;409(24):5423–5431.
39. Wilson J, Dixon S, Breysse P, et al Housing and allergens: a pooled analysis of nine US studies. Environ Res. 2010;110:189–198.
health and safety; healthy homes; healthy housing; home environment; housing; intervention; program evaluation