Economic Shocks and Public Health Protections in US Metropolitan Areas

Abstract

Objectives. We examined public health system responses to economic shocks using longitudinal observations of public health activities implemented in US metropolitan areas from 1998 to 2012.

Methods. The National Longitudinal Survey of Public Health Systems collected data on the implementation of 20 core public health activities in a nationally representative cohort of 280 metropolitan areas in 1998, 2006, and 2012. We used generalized estimating equations to estimate how local economic shocks relate to the scope of activities implemented in communities, the mix of organizations performing them, and perceptions of the effectiveness of activities.

Results. Public health activities fell by nearly 5% in the average community between 2006 and 2012, with the bottom quintile of communities losing nearly 25% of their activities. Local public health delivery fell most sharply among communities experiencing the largest increases in unemployment and the largest reductions in governmental public health spending.

Conclusions. Federal resources and private sector contributions failed to avert reductions in local public health protections during the recession. New financing mechanisms may be necessary to ensure equitable public health protections during economic downturns.

The large and growing economic burden of preventable diseases and injuries in the United States raises important questions about the adequacy of the nation’s public health protections. These protections include activities to monitor community health status, investigate and control disease outbreaks, educate the public about health risks and prevention strategies, prepare for and respond to natural disasters and other large-scale health emergencies, and enforce laws and regulations designed to protect health such as those concerning tobacco exposure, food and water safety, and air quality.¹ Although most of the $2.8 trillion spent annually on health care in the United States is attributable to preventable health conditions,² less than 5% of these expenditures support public health strategies that are designed to prevent and control disease and injury.^3,4 Limited investments in public health have been implicated as contributing to the higher rates of preventable diseases and premature mortality experienced in the United States compared with other high-income countries.^5,6

To help address this problem, the Patient Protection and Affordable Care Act of 2010 (ACA) created the Prevention and Public Health Fund and authorized as much as $15 billion in new federal spending over 10 years to support public health activities implemented primarily at state and local levels.⁷ Over time, however, federal policymakers have redirected large shares of fund resources to support other components of health reform implementation, such as the expanded training of primary care physicians, a temporary delay in scheduled Medicare physician payment cuts, and outreach and enrollment support for health insurance exchanges.^7,8

At the same time, the 2008 economic recession and subsequent slow pace of recovery have precipitated large and sustained reductions in state and local spending on public health activities. An estimated 55 000 public health workers were cut from state and local government workforces between 2008 and 2012.^9,10 Correspondingly, in 2011 the National Health Expenditure Accounts program recorded the first year-over-year reduction in total governmental public health spending since the federal government began tracking this spending category in 1980.³ The expiration of federal stimulus spending, which helped offset state and local public health spending cuts in 2009 and 2010, contributed to the magnitude and timing of this reduction. Most recently, the federal budget sequester of 2013 has placed additional downward pressure on public health resources, with significant reductions in federal transfers to state and local governments for public health activities. Taken together, the ACA Prevention and Public Health Fund diversions, stimulus expirations, and budget sequestration may have blunted the ability of federal outlays to moderate recessionary contractions in state and local public health spending.

These recent economic shocks raise the question of whether and how they have affected the implementation of public health protections in local communities. Government agencies may respond to funding reductions partly by improving operational efficiencies, such as by shifting to lower cost labor and technology or by enhancing the targeting and tailoring of programs to populations at greatest risk. Agencies may also reduce the scope and scale of public health activities they implement, causing a decline in the intensive margin of public health delivery as described in economic models of production.¹¹ As a third possible response, agencies may increase their reliance on the contributions of nongovernmental organizations in performing selected public health activities, thereby seeking to expand the extensive margin of public health delivery by bringing new resources into the production process. Indeed, several provisions of the ACA have created new incentives for the private provision of public health activities, such as enhanced community benefit requirements for tax-exempt nonprofit hospitals, minimum medical loss ratio regulations for health insurers that include credits for prevention and health promotion spending, and new federal tax incentives for employers to invest in wellness and health promotion programs.⁷ In theory, these policies may allow governmental agencies to play larger roles in mobilizing and coordinating the public health activities of nongovernmental organizations instead of performing some activities directly, resulting in an expanded extensive margin. Whether economic shocks leave communities more vulnerable to disease and injury depends critically on changes to the intensive and extensive margins of public health delivery. To explore these possibilities, we analyzed longitudinal observations of public health activities implemented in US metropolitan areas during 1998–2012.

METHODS

The National Longitudinal Survey of Public Health Systems has been used since 1998 to monitor local variation and change in the delivery of public health services across the United States.^12–20 The survey follows a nationally representative cohort of US communities using a validated questionnaire administered to the director of the local governmental public health agency in each community. The survey instrument asks these local officials to report information about a set of 20 public health activities that were identified by a consensus panel of public health professionals, scientists, and federal officials as providing important health protections (Table 1).^21–25 These activities were originally developed as indicators of 3 core functions of public health identified in the Institute of Medicine’s (IOM’s) 1988 consensus report on the US public health system¹ and were used to monitor progress toward the US Department of Health and Human Services’ Healthy People 2000²⁶ objective of reaching at least 90% of the US population with these 3 functions: (1) assessment of community health needs and risks, (2) development of policies and plans to address health needs, and (3) assurance that needed health services and supports are implemented in the community.^22–25 These 20 activities also correspond closely to the 10 Essential Public Health Services subsequently developed by the US Department of Health and Human Services in 1994²⁷ and with the set of foundational public health capabilities articulated more recently by IOM⁵ and a national expert consensus panel.²⁸ Although not a comprehensive inventory of public health protections, these 20 activities serve as a valuable screening tool for characterizing the breadth of public health work performed within communities.

TABLE 1—

Availability of Recommended Public Health Activities in Metropolitan Areas: National Longitudinal Survey of Public Health Systems, United States, 1998–2012

Public Health Activity	1998, %	2006, %	2012, %	% Change, 2006–2012
1. Community health needs assessment	71.5	77.5	72.6	−6.4*
2. Behavioral risk factor surveillance	45.8	70.2	73.9	5.2*
3. Adverse health events investigation	98.6	97.9	99.6	1.7
4. Public health laboratory testing services	96.3	97.0	99.2	2.2
5. Analysis of health status and health determinants	61.3	73.2	63.5	−13.3*
6. Analysis of preventive services utilization	28.4	26.1	33.2	27.3*
7. Health information provision to elected officials	80.9	90.1	87.1	−3.3
8. Health information provision to the public	75.4	88.8	80.9	−8.9*
9. Health information provision to the media	75.2	88.4	87.1	−1.4
10. Prioritization of community health needs	66.1	71.7	66.8	−6.8*
11. Community participation in health improvement planning	41.5	50.6	49.8	−1.7
12. Development of community health improvement plan	81.9	86.7	69.7	−19.6*
13. Resource allocation to implement the community health plan	26.2	37.3	27.8	−25.5*
14. Policy development to implement the community health plan	48.6	51.9	49.0	−5.7
15. Communication network of health-related organizations	78.8	87.2	89.6	2.8
16. Strategies to enhance access to needed health services	75.6	68.7	60.6	−11.8*
17. Implementation of legally mandated public health activities	91.4	92.3	89.2	−3.3*
18. Evaluation of public health programs and services	34.7	37.5	33.2	−11.5*
19. Evaluation of local public health agency capacity and performance	56.3	56.2	55.2	−1.8
20. Implementation of quality improvement processes	47.3	50.4	42.7	−15.3*
Composite for assessment activities (1–6)	66.7	73.8	73.7	−0.2
Composite for policy development activities (7–15)	60.2	68.1	62.8	−7.7*
Composite for assurance activities (16–20)	64.4	69.0	64.1	−7.0*
Composite for all activities (1–20)	63.8	70.2	66.9	−4.7*

Note. All measures reported by the local public health official in each jurisdiction.

Source. Authors’ analysis of data from the National Longitudinal Survey of Public Health Systems, 1998–2012.^13,19,20

*P < .05.

For each of the 20 activities, the survey instrument asks local public health officials to report on 4 dimensions of implementation: (1) whether the activity is performed in the community served by their public health agency (availability); (2) which types of organizations are involved in performing the activity (extensive margin), using a defined checklist of 15 organization types; (3) the proportion of effort contributed by the local governmental public health agency in performing the activity (intensive margin), measured using a 5-point Likert scale ranging from “none” to “all”; and (4) the perceived effectiveness of the activity in addressing community needs and risks, also measured using a 5-point Likert-type scale ranging from “not at all” to “fully addresses need.”^13,19 Composite measures of availability, intensive margin, and perceived effectiveness are constructed by computing a weighted average of the 20 activity measures, with activity weights defined so that each of the 3 IOM core functions receives equal weight in each composite measure. Two of these composite measures (intensive margin and perceived effectiveness) are constructed as weighted averages of Likert-type measures. Applying parametric statistics to Likert-type ordinal measures is sometimes not advisable because of violations of distributional assumptions, but the response sets used in this survey instrument were designed with numeric anchor points to support approximations to an interval scale. Parametric statistics used with these types of measures are robust to violations of distributional assumptions that result from treating Likert variables as interval measures, particularly in relatively large samples.²⁹

To construct the composite measure of extensive margin, we used the specification for network density derived from social network analysis theory. Specifically, we defined a link between any 2 organizations that both contributed to the same public health activity in the same community, and then we summed all links for all 20 activities in the community and divided by the total possible number of links.³⁰ This specification provides a measure of connectedness among organizations that contribute to public health activities, where the degree of connectedness between any pair of organizations is determined by the number of activities to which they jointly contribute. The instrument was developed and validated through a series of studies sponsored by the Centers for Disease Control and Prevention during the 1990s.^21–25

The National Longitudinal Survey of Public Health Systems panel followed a national cohort of metropolitan communities that were selected for the first wave of the survey in 1998.¹³ This cohort was defined by identifying all local governmental public health agencies in the United States that served jurisdictions containing at least 100 000 residents (n = 497). Each community is served by a single local public health agency, and there are no overlapping jurisdictions in this cohort. Three quarters of these agencies are units of county government and serve county-based jurisdictions; however, 9% of the agencies serve cities, and 16% serve special districts consisting of multiple counties and cities. Collectively, these jurisdictions are located predominantly in metropolitan areas (96%) and represent approximately 17% of all local public health jurisdictions in the United States, but they contain approximately 70% of the US population. The 2006 and 2012 waves of the National Longitudinal Survey of Public Health Systems resurveyed these same jurisdictions using the same instrument.^19,20 Response rates for each wave of the survey ranged from 68% to 73%, with no indication of systematic differences between responding and nonresponding agencies. Each wave of the National Longitudinal Survey of Public Health Systems is linked with data on the organizational and financial characteristics of local public health agencies collected by the periodic National Profile of Local Health Departments survey conducted by the National Association of County and City Health Officials in 1997, 2005, and 2013.^31,32 Additionally, survey data are linked with community-level demographic, health, and economic characteristics obtained from contemporaneous editions of the US Health Resources and Services Administration’s Area Health Resources File.

We used data from all 3 waves of the survey to detect changes over time in the local implementation of public health activities. We placed primary emphasis on changes that occurred between 2006 and 2012 because these periods occurred before and after the economic shocks associated with the Great Recession, federal stimulus, and initial ACA policy implementation. Recognizing that the magnitude of these shocks may have varied across states and communities, we used multivariate statistical models to estimate the changes in each composite measure of public health implementation attributable to local economic conditions, with a specific focus on the community’s unemployment rate, per capita income, and local public health agency expenditures per capita. These 3 indicators represent related but distinct dimensions of economic activity that are likely to vary in the magnitude and timing of their effects on the demand for and supply of public health activities. Unemployment and personal income track closely with state and local tax receipts and thereby influence governments’ fiscal capacity to support public health activities alongside other public services.³³ The relative influence of each indicator is likely to depend in part on the specific mix of tax instruments on which state and local governments rely, such as income and payroll taxes, sales taxes, and property taxes. Specifically, this mix will determine how sensitive governmental revenues are to changes in employment, income, and housing wealth. By contrast, public health agency expenditures reflect governmental fiscal effort: the degree to which local, state, and federal financial resources are allocated to local public health agencies. To isolate the effects of these local economic conditions on public health activities, our statistical models control for other community and institutional characteristics likely to influence the demand for and supply of public health activities, including local population size and density, racial and age composition of the population, hospital and physician availability, and the existence of governance and administrative structures that allow local versus state control of public health agency activities.

We estimate a separate statistical model for each of the 4 composite measures of public health implementation, using generalized estimating equations with a linear link function and an unstructured correlation specification for the error term. This specification accounts for the temporal correlation that exists among observations taken of the same agencies and communities over time, and it adjusts for the clustering of agencies and communities within the same state. All analyses treat the data as an unbalanced panel, allowing observations to be used even when data are missing for selected communities in selected years as a result of survey nonresponse or consolidation or deconsolidation among agencies. Results were similar but with larger standard errors when restricting analyses only to communities with nonmissing data in all 3 years of the panel.

RESULTS

The proportion of 20 recommended public health activities that are implemented in the average US community increased from 64% in 1998 to 70% in 2006, but then fell to 67% in 2012 (Table 1). This overall decrease in availability of nearly 5% between 2006 and 2012 (95% confidence interval [CI] = 4.4, 5.0) masks considerable variability across communities and activities. When communities are grouped into quintiles based on the composite availability measure in 2012, the bottom 20% of communities showed a 24.6% reduction in the availability of recommended activities between 2006 and 2012 (95% CI = 14.4, 34.9), and the top quintile showed a 22.9% increase (95% CI = 12.5, 33.3). Several public health activities remained nearly universally available during the period of study, including the investigation of adverse health events (available in 99.6% of communities), public health laboratory testing (99.2% availability), and the implementation of legally mandated public health services (89.2% availability) such as vital records collection and notifiable disease reporting.

We observed statistically significant reductions in availability for 9 of the 20 public health activities, whereas we observed increases in availability for only 2 activities, and 9 activities remained relatively stable. The largest reductions in availability between 2006 and 2012 occurred for activities that involved developing plans to improve community health (19.6% reduction), allocating resources on the basis of these plans (25.5% reduction), and implementing quality improvement processes for public health programs (15.3% reduction). Overall, policy development activities and assurance activities experienced significantly larger reductions in availability than assessment activities (P < .05).

Organizational Contributions and Perceived Effectiveness

The types of organizations that perform public health activities narrowed between 2006 and 2012—indicating an overall reduction in the extensive margin of public health delivery—with some of the largest reductions occurring among government agencies. Local governmental public health agencies performed 62% of the recommended public health activities in the average community in 2012, down from 67% in 2006 (Table 2). Participation by other local government agencies declined by 48.1%, federal government participation declined by 28.0%, and other state government participation fell by 21.0%. Participation by hospitals, health insurers, and community health centers showed more modest reductions than those reported for government agencies.

TABLE 2—

Participation by Selected Organizations in Implementing Recommended Public Health Activities: National Longitudinal Survey of Public Health Systems, United States, 1998–2012

Measure	1998, %	2006, %	2012, %	% Change, 2006–2012
Proportion of activities in which each organization participates (extensive margin)a
Local public health agencies	60.7	66.5	62.0	−6.8*
Other local government agencies	31.8	50.8	26.3	−48.1**
State public health agencies	46.0	45.3	36.4	−19.6**
Other state government agencies	17.2	16.4	13.0	−21.0*
Federal government agencies	7.0	12.0	8.7	−28.0*
Hospitals	37.3	41.1	39.3	−4.5*
Physician practices	20.2	24.1	19.5	−19.0*
Community health centers	12.4	28.6	26.9	−5.8
Health insurers	8.6	10.0	9.8	−2.5
Employers and businesses	25.5	16.9	13.4	−20.7*
Schools	30.7	27.6	24.9	−10.0*
Universities and colleges	15.6	21.6	21.2	−1.7
Faith-based organizations	24.0	19.2	15.7	−18.3*
Other nonprofit organizations	31.9	34.2	31.6	−7.8
Other organizations	8.5	8.8	5.4	−38.2
Composite extensiveness: all organizations	21.0	28.2	22.6	−19.9
Intensity of local public health agency participation (intensive margin)b
Assessment activities	38.4	40.2	37.7	−6.2
Policy development activities	34.0	39.3	34.2	−13.1*
Assurance activities	40.3	39.7	36.8	−7.3*
Composite intensity: all activities	37.6	39.7	36.2	−8.8*

Source. Authors’ analysis of data from the National Longitudinal Survey of Public Health Systems, 1998–2012.

^aPercentages indicate the proportion of 20 activities in which each type of organization participates.

^bPercentages indicate the proportion of 20 activities in which the local public health agency provides most or all of the effort to implement activity.

*P < .05; **P < .01.

Regarding the intensive margin of public health delivery, local officials reported significant reductions in the activities for which the local governmental public health agency provided all or most of the effort necessary to implement the activity (Table 2). The largest reductions in this intensive margin between 2006 and 2012 were reported among activities involving community health plan development (29.3% reduction), allocation of resources to the community health plan (41.3% reduction), and the provision of or linkage to health services (17.0% reduction). This last activity reflected the continuation of a larger downward trend over the entire study period from 1998 to 2012 in public health agency involvement in clinical services provision. Across all 20 activities, officials reported an average reduction in the intensive margin of 8.8% between 2006 and 2012, with larger reductions for policy development activities than for assessment and assurance activities (P < .05).

Perceptions of the effectiveness of public health activities in addressing community needs and risks improved significantly for 4 of the activities measured and declined significantly for 9 other activities from 2006 to 2012 (Table 3). All 4 of the activities receiving improved ratings of perceived effectiveness clustered within the functional domain of community health assessment. By contrast, the ratings of policy development activities declined by an average of 9.0%, and the ratings of assurance activities declined by an average of 7.8%.

TABLE 3—

Perceived Effectiveness of Recommended Public Health Activities in Metropolitan Areas: National Longitudinal Survey of Public Health Systems, United States, 1998–2012

Public Health Activity	1998, %	2006, %	2012, %	% Change 2006–2012
1. Community health needs assessment	35.3	50.1	53.7	7.2**
2. Behavioral risk factor surveillance	21.0	44.1	48.3	9.5**
3. Adverse health events investigation	75.1	83.7	86.6	3.4**
4. Public health laboratory testing services	72.9	78.7	80.3	2.0
5. Analysis of health status and health determinants	29.4	43.1	39.4	−8.6**
6. Analysis of preventive services utilization	12.1	14.7	19.6	33.0**
7. Health information provision to elected officials	37.7	57.3	54.0	−5.7**
8. Health information provision to the public	32.5	55.7	49.6	−11.0**
9. Health information provision to the media	39.5	63.2	60.7	−4.0*
10. Prioritization of community health needs	33.8	46.4	46.6	0.5
11. Community participation in health improvement planning	16.2	29.8	29.5	−1.2
12. Development of community health improvement plan	34.7	53.2	39.3	−26.1**
13. Resource allocation to implement the community health plan	10.5	20.3	14.7	−27.4**
14. Policy development to implement the community health plan	18.4	26.6	23.4	−11.9**
15. Communication network of health-related organizations	42.3	58.7	58.6	−0.1
16. Strategies to link people to needed health services	35.9	39.8	34.4	−13.5**
17. Implementation of legally mandated public health activities	—	—	—	—
18. Evaluation of public health programs and services	15.9	19.1	17.9	−6.1**
19. Evaluation of local public health agency capacity and performance	31.3	37.2	37.7	1.1
20. Implementation of quality improvement processes	21.6	27.8	24.0	−13.8**
Composite for assessment activities (1–6)	40.8	52.5	54.6	4.1*
Composite for policy development activities (7–15)	27.5	41.8	38.0	−9.0**
Composite for assurance activities (16–20)	37.7	40.6	37.4	−7.8**
Composite for all activities (1–20)	35.4	44.9	43.4	−3.5

Source. Authors’ analysis of data from the National Longitudinal Survey of Public Health Systems, 1998–2012.

Note. Percentages indicate the proportion of communities in which the activity is perceived to meet most or all of the community need for the activity. All measures reported by the local public health official in each jurisdiction.

*P < .05; **P < .01.

Effects of Economic Shocks

Multivariate estimates confirmed that local public health delivery fell most sharply among communities that experienced the largest reductions in public health agency spending and household income and the largest increases in unemployment during the economic recession (Table 4). A 10% reduction in local public health agency expenditures per capita was associated with statistically significant declines in all 4 measures of public health delivery: a 31 percentage point reduction in the availability of public health activities, a 36 percentage point reduction in the intensive margin of activities contributed by local public health agencies, a 6 percentage point reduction in the extensive margin of activities contributed by other organizations, and a 29 percentage point reduction in the perceived effectiveness of public health activities.

TABLE 4—

Generalized Estimating Equation Estimates of Factors Associated With Implementation of Recommended Public Health Activities: National Longitudinal Survey of Public Health Systems, United States, 1998–2012

	Dependent Variablea
Independent Variable (Scale)	Availability, b (95% CI)	Extensive Margin, b (95% CI)	Intensive Margin, b (95% CI)	Perceived Effectiveness, b (95% CI)
Public health expenditures per capita, log $	3.13** (1.11, 5.15)	0.56* (0.01, 1.11)	3.58** (2.01, 5.14)	2.89** (1.18, 4.61)
Unemployment rate, log %	−4.74* (−9.04, −0.45)	−1.99** (−3.16, −0.81)	−3.73* (−7.09, −0.37)	−3.01 (−6.73, 0.70)
Household income per capita, log $	6.44* (0.12, 12.76)	−0.04 (−1.78, 1.69)	−2.19 (−7.12, 2.73)	13.47** (8.04, 18.91)
Population size, log	2.78* (0.53, 5.03)	0.55 (−0.05, 1.15)	2.71** (0.99, 4.44)	1.54* (−0.34, 3.42)
Population per square mile, 10 000s	0.24 (−4.05, 4.53)	−0.21 (−1.37, 0.95)	3.32* (0.02, 6.62)	1.67 (−1.94, 5.27)
Board of health existsb	2.89 (−1.23, 7.01)	0.79 (−0.33, 1.92)	0.63 (−2.58, 3.84)	1.29 (−2.25, 4.83)
Centralized public health agencyb	−2.05 (−7.72, 3.63)	−0.50 (−2.03, 1.03)	−5.64* (−10.02, −1.25)	−1.57 (−6.39, 3.24)
Hospital beds per 100 000 residents	−0.15 (−1.29, 0.99)	−0.15 (−0.46, 0.16)	−0.14 (−1.02, 0.74)	0.32 (−0.64, 1.29)
Physicians per 100 000 residents	−0.45 (−1.72, 0.82)	−0.04 (−0.38, 0.30)	−0.24 (−1.22, 0.73)	−0.95 (−2.01, 0.12)
% of population non-White	−0.13* (−0.26, 0.00)	−0.02 (−0.05, 0.02)	−0.13** (−0.23, −0.03)	−0.10 (−0.21, 0.01)
% of population aged ≥ 65 y	0.42 (−0.09, 0.93)	0.14* (0.01, 0.28)	0.38 (−0.01, 0.78)	0.29 (−0.15, 0.72)
Constant	−39.20	7.94	20.38	−12.21
Scale parameter	304.04	22.93	184.78	225.23
No. of observations	590	590	590	590
No. of groups	234	234	234	234
Wald χ211	39.24	29.82	59.64	65.91
Probability > χ2	0.00	0.00	0.00	0.00

Note. CI = confidence interval. Parameter estimates and confidence intervals are from generalized estimating equation models using a linear link function and unstructured correlation specification.

Source. Authors’ analysis of data from the National Longitudinal Survey of Public Health Systems, 1998–2012.

^aThe dependent variables are the composite measures of public health activities shown in Tables 1, 2, and 3.

^bExpressed as a binary term, with 0 meaning no and 1 meaning yes.

*P < .05; **P < .01.

Similarly, growth in the unemployment rate was associated with significant declines in the availability of public health activities, the intensive margin of activities contributed by local public health agencies, and the extensive margin of activities contributed by other organizations. Declines in household income per capita were associated with significant reductions in both the availability and the perceived effectiveness of public health activities, but were not related to other measures of public health delivery after adjusting for other characteristics. Taken together, the 3 indicators of local economic conditions accounted for between 52% and 71% of the explained variation in our measures of public health delivery from 1998 to 2012.

DISCUSSION

The economic downturn beginning in 2008 triggered significant reductions in the delivery of public health protections in the communities in which most persons in the United States reside. Significant reductions occurred in the scope of public health activities that are performed within communities, the range of organizations that participate in performing these activities, the intensity of contributions from local governmental public health agencies, and the perceived effectiveness of these activities in addressing community health needs. The incidence of these reductions is not surprising given the severity and duration of the economic downturn and the heavy reliance on state and local governmental financing mechanisms for implementing public health activities across the United States. Reassuringly, public health activities related to community health assessment and epidemiological investigation appeared relatively resistant to the economic downturn, and many of these activities continued to be performed at high levels.

These findings demonstrate a relatively robust capacity for the timely detection of and short-term response to emerging health threats such as communicable disease outbreaks and responses to large-scale health emergencies. However, public health activities related to policy development, planning, and implementation functions appeared to be among those hardest hit by recessionary pressures, raising concerns about local-level capacity to develop and mount effective, longer term responses to preventable diseases and injuries. Such public health activities often require sustained and consistent implementation periods over time to be most effective, suggesting that the disruptive effects of economic shocks may undermine their population-wide health impact. These findings reveal that the tendency to favor immediate over long-term needs—a phenomenon called “present bias” in behavioral economics³⁴—exists not only in individual health-related decision making but also at the community level and that public health officials face difficulties in counteracting this tendency.

The moderate magnitude of reductions in public health delivery that we observed, contrasted with the steep reductions in overall state and local government resources experienced during the recession, led to the conclusion that it could have been worse. One plausible explanation for the muted impact of the economic recession on public health delivery is the offsetting effect of federal stimulus spending and ACA funding. The American Recovery and Investment Act of 2009³⁵ triggered new federal spending on state and local public health activities through the Centers for Disease Control and Prevention’s Putting Prevention to Work Program from 2009 to 2012, and the ACA also generated new federal funding for public health activities over this period through Prevention and Public Health Fund programs such as the Community Transformation Grants and the National Public Health Improvement Initiative. Of course, some of the “new” federal public health spending generated through these programs was offset by cuts to existing federal public health programs during this period. Nevertheless, increases in federal public health spending are likely to have played an important role in moderating recession-induced cuts to public health delivery at the local level.

Additionally, local shifts in the extensive and intensive margins of public health delivery may have played a role in buffering public health protections from recessionary contractions. Overall, we found that both the range of organizations contributing to public health delivery and the intensity of contributions by governmental public health agencies declined significantly from 2006 to 2012 period. However, the contributions to public health activities made by health care delivery system stakeholders such as hospitals, health insurers, and community health centers proved to be much more recession resistant than were the contributions made by governmental agencies during this time period. These differentials in activities and organizational contributors appear likely to have muted the effects of the recession on public health delivery.

These changes result in a public health system that is more dependent on the health care delivery system today than it was before the recession. As a consequence, public health officials may face new opportunities for coordinating and integrating medical care and public health programs in ways that improve their collective reach, effectiveness, and efficiency. The Institute of Medicine recently identified promising strategies for integrating public health and primary care delivery, some of which are facilitated through new delivery models supported by the ACA such as accountable care organizations, patient-centered medical homes, and community health worker programs.³⁶ Our findings suggest that the local public health activities currently contributed by health care providers may provide a foundation for supporting expanded models of coordination and integration under ACA.

Public health assessment activities appeared remarkably resistant to recessionary forces, possibly reflecting several recent developments in national health policy. First, the voluntary national accreditation program for public health agencies, launched in 2012, requires agencies to conduct community health assessment and community health improvement planning activities every 5 years to be eligible for accreditation.^37–40 Second, the ACA created enhanced community benefit requirements for not-for-profit hospitals that receive federal tax exemptions, including the obligation to conduct a community health needs assessment every 3 years and to produce and implement a community health improvement plan in collaboration with local public health stakeholders.⁴¹ The combined effects of these policies may have led communities to prioritize assessment and planning activities over other types of public health activities when responding to economic shocks. The intent of these policies is to encourage assessment and planning activities that help communities achieve higher rates of success with implementing coordinated, multiorganizational strategies that improve health status on a community-wide basis.^12,42 Our results do not show evidence of improvements in this type of implementation, but it could be a longer term result of the sustained assessment activities that we observed between 2006 and 2012. To this end, the community benefit expenditures supported annually by not-for-profit hospitals represent a large resource base that could be tapped to support expanded implementation of public health activities, particularly in states and communities in which the demand for hospital charity care declines under ACA’s insurance coverage expansions.^41,43

This study includes several important limitations that should be kept in mind. First, our data were derived almost exclusively from metropolitan communities and may not reflect the dynamics of public health delivery experienced in small and rural settings. Second, our measures of public health delivery are based on the reports of local public health officials and may not fully capture the contributions of other organizations in the community, particularly when these contributions are implemented independently or when they target relatively small and less visible segments of the community population. Validation studies have confirmed that these officials provide reliable estimates of the total supply of public health activities in their jurisdictions,^21,22 but nevertheless our results may represent lower bound estimates of community-wide public health delivery.

Overall, the findings from this study demonstrate that federal spending offsets and local changes in the extensive and intensive margins of public health delivery are not sufficient to counteract the adverse impact of economic shocks on local public health delivery. The Great Recession of 2008 widened the gap between preventable disease burdens and public health protections in the communities in which most persons in the United States reside, and communities hit hardest by the recession suffered the largest losses in protection. The health and economic consequences of this gap will not be realized immediately, but their cumulative effects over time could be substantial. Recent research has shown that communities with lower spending on public health protections experience significantly higher rates of preventable mortality over time.⁴⁴ Heavy reliance on state and local government financing for public health activities makes many communities vulnerable to economic shocks that weaken health protections and thereby help to institutionalize health inequities based on socioeconomic status and geography.³³ Because public health activities are true public goods that benefit broad segments of the population over relatively long periods of time, the private sector may not face sufficient economic incentives to support them at socially optimal levels and to create stable financing mechanisms.⁴⁵ Experimentation with new, long-term financing instruments, such as social impact bonds, may help public health protections become more resilient to economic shocks.⁴⁶ However, the Institute of Medicine’s call for a larger federal role in financing public health activities⁵ may also prove necessary to ensure that these protections weather economic cycles and achieve more equitable coverage across the United States.