Abstract
Background
Healthcare delivery produces substantial emissions that contribute to climate change and harm human health. Patient perspectives on ethical dilemmas, such as tradeoffs between individual health choices and public health harms mediated by climate change, are unclear.
Methods
This cross-sectional survey randomly sampled adult patients across four US health systems to assess their perspectives on ethical dilemmas in climate change and healthcare delivery; results were compared to a previous nationwide survey of US-based physicians. The mailed survey was developed iteratively through pre-testing and was designed to detect a 15% difference in the proportion willing to limit treatment options because of environmental impact according to respondents’ perceived impact of climate change on their health. Secondary outcomes included physician responsibilities for healthcare sustainability and acceptability of environmentally motivated treatment tradeoffs.
Results
Between 11/2023 and 9/2024, 289 of 516 patient surveys and 304 of 529 physician surveys were delivered and returned, for response rates of 56.0% and 57.5%, respectively. Most patients (79.1%) believed that environmental factors impacted their medical conditions, and 36.3% reported a moderate-to-high health impact from climate change, while 5.2% reported speaking with their doctor about climate and health interactions a moderate amount or more. Similar proportions of patients (35.8%) and physicians (35.0%) agreed with reducing healthcare’s environmental impact even if it required limiting treatment options. Like physicians, patients’ perceived health impact (moderate-to-high versus low-to-no) was associated with willingness to place such limits (adjusted OR 1.85; 95% CI 1.01, 3.41). Most patients (77.1%) were willing to accept some reduction in the likelihood of treatment response if that treatment was less environmentally impactful; unlike physicians, this did not vary by health impact (adjusted OR 1.16; 95% CI 0.63, 2.20). Almost all patients (96.8%) reported that physicians should help make healthcare sustainable, and 64.7% thought this included changing clinical practices.
Conclusions
Many US patients and physicians recognize connections between health, healthcare delivery, and climate change, and accept environmentally motivated treatment tradeoffs, but do not discuss them in the clinic. Patient views largely parallel those of physicians, suggesting support for climate-informed medical practice and for incorporating environmental considerations into clinical decision-making.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12916-026-04656-8.
Keywords: Climate change, Health care delivery, Medical ethics, Patient attitudes, Environmental health, Surveys and questionnaires
Background
Healthcare in the United States (US) generates a substantial environmental footprint. Despite making up only 4% of the world’s population, US healthcare produces 20% of global healthcare emissions at more than twice the per-capita rate of similar countries [1–3]. These emissions lead to a loss of life similar in scale to chronic obstructive pulmonary disease or diabetes mellitus [4]. Over the past decade, health systems and governments have begun to address healthcare emissions through mitigation, adaptation, and resilience programs and policies. Most of these efforts have aimed to curb major emissions sources such as healthcare facilities and the supply chain. Evidence now shows that clinical practices also contribute substantially and directly to emissions while influencing facilities and the supply chain [5–9]. This raises new questions about how clinicians should approach their obligations to individual patients alongside these public health implications [10]. What responsibilities do clinicians have to consider environmental impact in their practice? How should they navigate situations where reducing environmental harm may require limiting certain treatment options?
We recently surveyed US physicians to explore their views on these ethical dilemmas and openness to climate-informed clinical care [11]. The perspectives of patients, however, remain unclear. Recent studies have assessed patient views on climate-related healthcare safety, counseling, and delivery. They found that patients were open to some aspects of climate-informed care but were concerned with limiting health choices [12–14]. These studies were performed outside the US, however, and did not compare patients views to those of physicians.
In this context, we conducted a multi-institutional survey of US adult patients to characterize their views on ethical dilemmas in climate change and healthcare delivery, and to compare their perspectives to those of US physicians. We posed three research questions: (1) how do US patients view the relationship between climate change, their health, and their physicians’ responsibilities; (2) would they accept limits on treatment options or efficacy reductions to reduce environmental impact, and do these vary by perceived health impact; and (3) how do these views compare with those of US physicians? We hypothesized that patients reporting greater climate-related health impacts would be more willing to accept environmentally motivated tradeoffs, that most patients would endorse a role for physicians in sustainable healthcare, and that patients’ willingness to accept environmentally motivated tradeoffs was similar to physicians.
Methods
Study design, objectives, and endpoints
We performed a cross-sectional survey of adult patients to assess their views on physician professional responsibilities for sustainable healthcare, interactions between individual and environmental health, and the potential impact of climate change on health equity. We define climate change as referring to the long-term warming and shifting of the global climate caused mainly by human activities, while environmental impact refers to the effect human activities have on natural systems like air, water, and ecosystems. The primary endpoint was the difference in the proportion of patients willing to limit treatment options of similar efficacy due to environmental impact, based on their self-reported health impact from climate change. Secondary endpoints were the proportions reporting physician responsibility for healthcare sustainability, acceptability of tradeoffs between individual treatment choice and environmental impact, and differences in patient versus physician viewpoints with respect to the other endpoints. The primary and secondary endpoint questions and response options are shown in Table 1. The methods and full results of our physician survey were previously reported elsewhere; physicians in that survey were randomly sampled from the US National Plan & Provider Enumeration System, stratified into primary and specialty care [11]. Physician results are included here exclusively as a comparator group.
Table 1.
Comparator group and endpoint questions and response options
| Question | Response Options | Groupings | |
|---|---|---|---|
| Within-Patient Comparator Groups | Climate change has affected my health | I don’t know | No-to-low impact |
| Not at all | |||
| A little | |||
| A moderate amount | Moderate-high impact | ||
| A great deal | |||
| Primary Endpoint | The environmental impact of the healthcare system should be reduced even if it requires limiting treatment options for individuals | Strongly Disagree | Disagree |
| Disagree | |||
| Agree | Agree | ||
| Strongly Agree | |||
| Secondary Endpoints | Based on the definition [of sustainable healthcare] above, should doctors support sustainable healthcare? | Yes | Yes |
| No | No | ||
| Pretend that you have high blood pressure. Medicines C and D are approved to lower your blood pressure. Medicine C has a 90% chance of working and is environmentally toxic. Medicine D has a ___ % chance of working and is not environmentally toxic. If the medicines work, they work equally well and have the same side effects and costs. If Medicine D does not work, you can switch to Medicine C without extra cost. How low could the missing value be for you to try Medicine D fist? | 90% | 90% | |
| 85% | < 90% | ||
| 80% | |||
| 70% | |||
| 60% | |||
| 50% or less |
Participants and recruitment
Patients were recruited from the University of Chicago, Mount Sinai, Dana-Farber Cancer Institute, and Brigham and Women’s Hospital health systems in Illinois, Indiana, New York, and Massachusetts, and New Hampshire. While these are academic health systems, the samples included patients across the main medical centers and their affiliated, community-based clinical sites. Samples of patients who met the following criteria were generated by health system registries: age 18 or older, preferred language of English or Spanish, at least one clinical visit within the last 12 months, no diagnosis of severe cognitive impairment, and a valid mailing address in the United States. Random sampling stratified by preferred language and race/ethnicity was performed within each hospital system to generate lists of patients; stratifications were used to enhance generalizability to historically under-surveyed demographic groups because of known associations between climate and health views and the selected demographics [15]. After this, eligibility was verified until a prespecified sample size was reached, as described below. Details of the sampling methods are described in the supplementary material (Additional File 1). Paper surveys were mailed to potential participants through a private courier service (Federal Express). Following best practices for maximizing response [16], surveys included a prepaid return envelope and a modest incentive ($20 gift cards); a follow-up letter with an electronic response option was sent after 2 weeks, and a follow-up phone call for non-responders was made 2 weeks after that. Approval for the study was obtained by the institutional review board of each participating institution.
Survey development
Our general approach to instrument construction and development has been described in detail elsewhere [11, 17]. Briefly, parallel survey instruments were constructed for patients and physicians based on the research questions outlined above; demographic questions were adapted from the American Community Survey and US Decennial Census [18] and climate change belief questions were adapted from Medical Consortium on Climate Change and Health surveys [19, 20]. The draft patient instrument was revised through formal testing with 27 patients across four focus groups (three English language, one Spanish language) [17] and then pretested by 3 lay individuals and a survey methodologist [21]. The final patient instrument included 27 questions in 7 sections (demographics, climate change views, climate change health impact, environment and health discussions, responsibilities, limitations, education) and is shown in the supplementary material (Additional File 1).
Statistical analysis
Survey responses from patient and physician cohorts were aligned for comparative analyses. Transformations of survey response options for analysis are outlined in Table 1 and the supplementary material (Additional File 1). To assess response associations with living in areas susceptible to climate change, place of residence was linked to the county-level Climate Vulnerability Index (CVI) [22], a validated measure that reflects baseline vulnerability and reduced resilience to climate change risks that impact health. CVI scores are normalized risk indices that range from 1 to 100, with 100 being areas most vulnerable to climate change; scores were analyzed grouped into quartiles and labeled from “Lowest Vulnerability” to “Highest Vulnerability.”
The survey was powered to detect an approximately 15% difference in respondent willingness to limit individual treatment options based on environmental impact among two groups defined by perceived climate-related health impact (no-to-low versus moderate-high impact). The survey question and response options that define these groups are shown in Table 1. A 15% difference was selected based on anticipated between-group differences seen in prior surveys of patient and clinician perspectives on climate and health [19, 23, 24], and because smaller effects would be unlikely to materially alter clinical decisions or institutional planning. The comparator groups were selected because there was a connection between personal health impact and willingness to make changes in healthcare delivery identified during qualitative focus group analysis [17]. Based on responses during survey development, we assumed unbalanced groups (0.60 low, 0.40 high) and response proportions (0.75 no, 0.25 yes), two-sided alpha 0.05, and 80% power, which led to a required minimum of 282 respondents. Assuming a response rate of 55% and 3% undeliverable, we mailed 528 patient surveys.
Descriptive statistics, including means and proportions with 95% confidence intervals (CI) and medians with interquartile ranges (IQR), were calculated for continuous and categorical variables. For categorical variables, X2 tests or Fisher’s exact tests were performed to assess bivariate associations by no-to-low versus moderate-high impact groups and by respondent type (patients and physicians). Independent t-tests and one-way ANOVA were used for continuous outcomes. Non-parametric alternatives were used when parametric test assumptions were violated. For primary and secondary endpoints, multivariable logistic regression models were used to assess associations between patients’ perception of climate change impact on personal health and key outcomes: willingness to limit treatment options based on environmental impact and acceptable reduction in treatment efficacy when choosing less environmentally harmful options.
Regression models were adjusted for potential confounders including age group, gender, race/ethnicity, and CVI score; these covariates were included because of known associations with general views on climate change [24] and/or because of concern for confounding. Models with aggregated race/ethnic grouping (as described in the supplementary material (Additional File 1)) are shown. Disaggregated models were also performed but are not reported as results were consistent. Separate models that included physician respondents were performed. Variance inflation factors were calculated to assess multicollinearity among predictor variables. Statistical significance was determined using a two-sided p-value threshold of 0.05. Analyses were of complete cases as item response missingness was < 2%. All analyses were conducted using RStudio with R version 4.3.2, and manuscript reporting was performed according to the Checklist for Reporting of Survey Studies reporting standards.
Results
Between November 17, 2023 and September 3, 2024, 289 of 516 patient surveys and 304 of 529 physician surveys were delivered and returned for a patient response rate of 56.0% and a physician response rate of 57.5%. Of the patient surveys, 12 (2.2%) were undeliverable and there were 10 (1.9%) opt-outs and 227 (44.0%) no responses. Of the 516 delivered patient surveys, 60 of 121 (49.6%) Spanish language and 229 of 395 (58.0%) English language surveys were returned (p = 0.10). Self-reported demographics of patient respondents are shown in Table 2 with comparisons to physician respondents; 62.8% of patients identified as female and 40.5% from a minoritized racial or ethnic group.
Table 2.
Respondent characteristics
| Characteristic |
Overall, N = 5931 |
Patient, N = 2891 |
Physician, N = 3041 |
p-value2 |
|---|---|---|---|---|
| Age in years | < 0.001 | |||
| 20–39 | 112 (19.3%) | 26 (9.2%) | 86 (29.1%) | |
| 40–59 | 237 (40.9%) | 96 (33.9%) | 141 (47.6%) | |
| 60–79 | 197 (34.0%) | 132 (46.6%) | 65 (22.0%) | |
| ≥ 80 | 33 (5.7%) | 29 (10.2%) | 4 (1.4%) | |
| N missing | 14 | 6 | 8 | |
| Gender | < 0.001 | |||
| Female | 308 (52.5%) | 181 (62.8%) | 127 (42.5%) | |
| Male | 279 (47.5%) | 107 (37.2%) | 172 (57.5%) | |
| N missing | 6 | 1 | 5 | |
| Hispanic, Latino, or of Spanish origin | < 0.001 | |||
| Yes | 94 (16.1%) | 74 (26.1%) | 20 (6.7%) | |
| No | 478 (82.0%) | 205 (72.4%) | 273 (91.0%) | |
| I choose not to answer | 11 (1.9%) | 4 (1.4%) | 7 (2.3%) | |
| N missing | 10 | 6 | 4 | |
| Race3 | ||||
| American Indian or Alaska Native | 3 (0.5%) | 2 (0.7%) | 1 (0.3%) | 0.6 |
| Asian Indian | 29 (4.9%) | 4 (1.4%) | 25 (8.2%) | < 0.001 |
| Black or African American | 56 (9.4%) | 45 (15.6%) | 11 (3.6%) | < 0.001 |
| Eastern Asian or Pacific Islander | 34 (5.7%) | 6 (2.1%) | 28 (9.2%) | < 0.001 |
| White | 405 (68.3%) | 186 (64.4%) | 219 (72.0%) | 0.045 |
| A race not listed | 32 (5.4%) | 23 (8.0%) | 9 (3.0%) | 0.007 |
| I choose not to answer | 35 (5.9%) | 23 (8.0%) | 12 (3.9%) | 0.038 |
| Minoritized racial-ethnic group | 0.001 | |||
| Minoritized group | 211 (35.6%) | 117 (40.5%) | 94 (30.9%) | |
| Non-Hispanic White | 346 (58.3%) | 148 (51.2%) | 198 (65.1%) | |
| I choose not to answer | 36 (6.1%) | 24 (8.3%) | 12 (3.9%) | |
| Region | < 0.001 | |||
| North Central | 146 (25.2%) | 67 (24.2%) | 79 (26.1%) | |
| Northeast | 258 (44.5%) | 204 (73.6%) | 54 (17.8%) | |
| South | 105 (18.1%) | 4 (1.4%) | 101 (33.3%) | |
| West | 71 (12.2%) | 2 (0.7%) | 69 (22.8%) | |
| N missing | 13 | 12 | 1 | |
| Home/Practice Location CVI Score4 | < 0.001 | |||
| Highest Vulnerability | 42 (7.1%) | 8 (2.8%) | 34 (11.2%) | |
| Moderate-High | 213 (35.9%) | 131 (45.3%) | 82 (27.0%) | |
| Low-Moderate | 204 (34.4%) | 106 (36.7%) | 98 (32.2%) | |
| Lowest Vulnerability | 134 (22.6%) | 44 (15.2%) | 90 (29.6%) | |
| Highest degree or level of school completed5 | ||||
| Associates degree or less | – | 112 (40.1%) | – | – |
| Bachelor’s degree | – | 80 (28.7%) | – | |
| Advanced degree | – | 87 (31.2%) | – | |
| N missing | – | 10 | – | |
| Annual household income5 | ||||
| Less than $50,000 | – | 78 (32.1%) | – | – |
| $50,000 to $150,000 | – | 98 (40.3%) | – | |
| $150,000 or more | – | 67 (27.6%) | – | |
| N missing | – | 46 | – | |
| Climate change is happening | 0.5 | |||
| Yes | 552 (94.5%) | 264 (93.6%) | 288 (95.4%) | |
| No | 9 (1.5%) | 4 (1.4%) | 5 (1.7%) | |
| I don’t know | 23 (3.9%) | 14 (5.0%) | 9 (3.0%) | |
| N missing | 9 | 7 | 2 | |
| Climate change cause | 0.030 | |||
| Entirely or mostly human activities | 393 (68.0%) | 174 (62.6%) | 219 (73.0%) | |
| Equally human activities and natural changes | 128 (22.1%) | 75 (27.0%) | 53 (17.7%) | |
| Entirely or mostly by natural changes | 25 (4.3%) | 11 (4.0%) | 14 (4.7%) | |
| I don’t know/do not think climate change is occurring | 32 (5.5%) | 18 (6.5%) | 14 (4.7%) | |
| N missing | 15 | 11 | 4 | |
| Climate change has affected my health/my patients’ health | 0.002 | |||
| Moderate or A great deal | 215 (36.9%) | 102 (36.3%) | 113 (37.4%) | |
| Not at all or A little | 271 (46.5%) | 117 (41.6%) | 154 (51.0%) | |
| I don’t know/do not think climate change is occurring | 97 (16.6%) | 62 (22.1%) | 35 (11.6%) | |
| Climate change has affected the health of my community.5 | ||||
| Low Community Impact | – | 127 (48.5%) | – | – |
| High Community Impact | – | 135 (51.5%) | – | |
| N missing | – | 27 | – | |
| In the next 10 years, climate change will affect the health of my community.5 | ||||
| Low Community Impact | – | 74 (28.4%) | – | – |
| High Community Impact | – | 187 (71.6%) | – | |
| N missing | – | 28 | – | |
1n (%)
2Pearson’s Chi-squared test; Fisher’s exact test
3Non-mutually exclusive response categories, testing performed within each group
4CVI climate vulnerability index
5Not asked in physician survey
A high proportion of patients (93.6%) thought climate change was occurring; among those, 62.6% thought it was caused entirely or mostly by human activities, and 75.9% thought it would have a greater impact on those with less healthcare access. Most patients (79.1%) reported that the environment impacted their medical conditions and that they wanted to reduce the environmental impact of their healthcare (75.5%), while 36.3% reported a moderate or great deal of impact from climate change on their health (which constituted the moderate-high comparator group). Only 30.7% reported knowing how to reduce the environmental impact of their healthcare, and 5.2% reported speaking with their doctor about the interaction of the environment with their health a moderate amount or more. These results are shown in table format, along with complete responses to questions about climate change’s perceived impact on patient health, in Table 1 and the supplementary material (Additional File 1).
Nearly half of the 75.5% of patients who reported wanting to reduce the environmental impact of their healthcare agreed with reducing its impact even if it required limiting treatment options (35.8% of all patients). This was similar to physicians, where 35.0% agreed with limiting treatment options (p = 0.80). Physicians reporting a moderate-high impact of climate change on their patients’ health were more likely to accept limiting treatment options than physicians reporting no-to-low impact (adjusted odds ratio [OR] 4.34; 95% CI 2.51, 7.63). Patients’ perceived degree of personal health impact was also associated with accepting such limitations, though the degree of the association was more modest (adjusted OR 1.85; 95% CI 1.01, 3.41; Table 3).
Table 3.
Multivariable logistic regression of willingness to limiting treatment options based on their environmental impact (N = 225)
| Characteristic | OR1 | 95% CI1 | p-value |
|---|---|---|---|
| Climate change perceived personal or patient impact | |||
| No-to-Low Health Impact | — | — | |
| Moderate-High Health Impact | 1.85 | 1.01, 3.41 | 0.047 |
| Age in years | |||
| 20–39 | — | — | |
| 40–59 | 0.79 | 0.28, 2.31 | 0.7 |
| 60–79 | 1.17 | 0.40, 3.55 | 0.8 |
| ≥ 80 | 1.74 | 0.44, 7.04 | 0.4 |
| Gender | |||
| Female | — | — | |
| Male | 1.32 | 0.70, 2.49 | 0.4 |
| Minoritized racial-ethnic group | |||
| Minoritized group | — | — | |
| Non-Hispanic White | 0.57 | 0.28, 1.16 | 0.12 |
| I choose not to answer | 1.19 | 0.36, 3.91 | 0.8 |
| Highest degree | |||
| Associates degree or Less | — | — | |
| Bachelor’s degree | 0.92 | 0.39, 2.19 | 0.9 |
| Advanced degree | 1.14 | 0.45, 2.95 | 0.8 |
| Home Location CVI ScoreI | |||
| Lowest Vulnerability | — | — | |
| Low-Moderate Vulnerability | 0.43 | 0.16, 1.12 | 0.08 |
| Moderate-High Vulnerability | 1.08 | 0.43, 2.76 | 0.9 |
| Highest Vulnerability | 0.78 | 0.11, 5.11 | 0.8 |
| Household Income | |||
| Less than $50,000 | — | — | |
| $50,000 to $150,000 | 0.98 | 0.40, 2.40 | > 0.9 |
| $150,000 or more | 0.97 | 0.32, 2.93 | > 0.9 |
IOR odds ratio, CI confidence interval, CVI climate vulnerability index
When two medicines equal in benefit, risk, and cost but differing in environmental impact were presented, 54.6% of patients reported they would want to be recommended the option with less environmental harm but still be informed about the alternative. Nearly a quarter (24.6%) preferred not mentioning the more environmentally harmful option at all, while 20.8% would want the decision to be left to them. This was different from physicians, where these proportions were 74.3%, 12.3%, and 13.3%, respectively (p < 0.001; Fig. 1). In a similar scenario, most patients believed the more environmentally impactful medicine should remain on the market and accompanied by a warning for patients and physicians (57.0%) or for physicians only (11.5%); more than a quarter (27.6%) preferred the medicine be removed from the market and 3.9% preferred no changes. While similar proportions of physicians preferred a physician and patient alert or no changes (57.0% and 8.6%), fewer preferred market removal (7.3%) and more the physician-only alert (26.6%).
Fig. 1.
Reponses to the question shown among patients (top panel) and for patients and physicians stratified by health impact categories (bottom panel)
When presented with a clinical scenario that asked patients how much of a reduction in expected treatment response was allowable for them to accept a less environmentally toxic antihypertensive medicine (Fig. 2), only 22.9% did not allow any reduction in efficacy while 43.9% allowed a 1–10% reduction in efficacy and 33.2% a more than 10% reduction. These results were similar to physicians (p = 0.60). Unlike physicians, where those in the moderate-high impact group were more willing to initially attempt treatment with a medicine with lower efficacy than those in the no-to-low impact category (adjusted OR 3.01, 95% CI 1.59, 6.04), there was no difference between the analogous patient groups (adjusted OR 1.16; 95% CI 0.63, 2.20). Full model results are shown in Table 4, and complete results of perspectives on environmentally related limitations of care are shown in the supplementary material (Additional File 1).
Fig. 2.
Responses to the question shown among patients (top panel) and for patients and physicians stratified by health impact categories (bottom panel)
Table 4.
Multivariable logistic regression of willingness to accept any reduction in treatment efficacy to reduce environmental impact. Positive odds indicate increased likelihood to accept a reduction in treatment efficacy. Covariates that were not collected from both groups were not included (e.g., household income for patients, specialty for physicians)
| Characteristic | Patients, N = 270 | Physician, N = 287 |
||||
|---|---|---|---|---|---|---|
| OR1 | 95% CI1 | p-value | OR1 | 95% CI1 | p-value | |
| Climate change perceived personal or patient impact | ||||||
| Low Personal/Patient Health Impact | — | — | — | — | ||
| High Personal/Patient Health Impact | 1.16 | 0.63, 2.20 | 0.6 | 3.01 | 1.59, 6.04 | 0.001 |
| Age in years | ||||||
| 20–39 | — | — | — | — | ||
| 40–59 | 1.34 | 0.43, 3.77 | 0.6 | 0.88 | 0.44, 1.71 | 0.7 |
| 60–79 | 1.03 | 0.34, 2.78 | > 0.9 | 1.10 | 0.49, 2.50 | 0.8 |
| ≥ 80 | 0.55 | 0.15, 1.87 | 0.3 | 0.12 | 0.01, 1.15 | 0.091 |
| Gender | ||||||
| Female | — | — | — | — | ||
| Male | 1.12 | 0.60, 2.12 | 0.7 | 0.57 | 0.31, 1.03 | 0.065 |
| Minoritized racial-ethnic group | ||||||
| Minoritized group | — | — | — | — | ||
| Non-Hispanic White | 0.99 | 0.51, 1.91 | > 0.9 | 0.87 | 0.44, 1.66 | 0.7 |
| I choose not to answer | 1.39 | 0.44, 5.37 | 0.6 | 0.31 | 0.07, 1.49 | 0.13 |
| Home/Practice Location CVI Score | ||||||
| Lowest Vulnerability | — | — | — | — | ||
| Low-Moderate | 0.83 | 0.32, 2.03 | 0.7 | 0.67 | 0.32, 1.38 | 0.3 |
| Moderate-High | 1.08 | 0.41, 2.67 | 0.9 | 0.75 | 0.34, 1.63 | 0.5 |
| Highest Vulnerability | 0.40 | 0.07, 2.39 | 0.3 | 0.46 | 0.18, 1.21 | 0.11 |
1OR odds ratio, CI confidence interval
High proportions of patients (96.8%) and physicians (89.7%; p < 0.001) reported that doctors should support sustainable healthcare. The extent to which each group viewed different actions as ones that physicians should take part in to fulfill that responsibility is shown in Fig. 3 and the supplementary material (Additional File 1). For both groups, the most commonly supported actions were physicians participating in hospital sustainability programs (patients: 71.6%, physicians: 72.0%) and changing clinical practices (patients: 64.7%, physicians: 70.7%). Fewer physicians agreed with changing research practices, meeting with government officials, and posting on social media, and more physicians agreed with reducing their work-related care travel (all p < 0.05).
Fig. 3.
Responses to the question shown among patients and physicians
Discussion
In this multi-institutional survey, US patients reported interactions between climate change and the environment with their health. They also expressed a willingness to accept certain tradeoffs in their clinical care to reduce environmental impact, and they considered physicians to have clinical responsibilities related to climate and health. Compared with physician respondents from a parallel survey [11], there was more consensus among patients on ethical issues at the intersection of climate and healthcare delivery, with minor differences regarding how best to approach climate-informed informed consent and decision-making. Like physicians, patients who reported experiencing more climate-related health impacts were more willing to accept environmentally motivated limits on care. Together, these data suggest that climate-informed medical practice may be acceptable to US patients as well as consistent with emerging professional norms. These data also highlight the need for climate-informed medical education so that physicians can adopt clinical practices that meet those expectations and norms.
The environmental footprint of US healthcare has grown substantially over the last three decades and now requires a per capita energy footprint two times greater than similar countries [3]. Recognition of this impact has led to interest in understanding the intersection between climate change and health equity, policy, and economics, among others [2, 3, 25]. Evidence about how clinical care impacts climate change is now accumulating rapidly [5–9, 26, 27], but our findings show that the adoption of climate-informed medical practice and decision-making remains limited and that key ethical dilemmas remain unaddressed.
Two factors might explain this disconnect. First, the evidence base for climate-informed care is still in development. While there is broad knowledge about general risks and harms (e.g., exposure to heatwaves increases renal and cardiovascular morbidity and overall mortality) [28, 29], the translation of these data into evidence-based action is much more limited [30]. Thus, while climate-informed practice to discuss climate-related risks with patients can be pursued generally, more research is needed to identify which specific patient are most at risk, for which outcomes, when, and what actions might mitigate that risk. There are similarly incomplete data on how clinical decisions propagate or mitigate climate change. In particular situations, evidence has been generated (e.g., inhaler choices for asthma [26, 27], the use of telehealth) [7, 31], but in many instances, the environmental impact of a clinical choice is unknown.
Second, given how separated US medical practice has been from environmental concerns, physicians and many other health professionals may not have received sufficient training about climate and health. Only recently have US medical schools included climate change in their curricula [32, 33] and continuing medical education is even less developed. Though several programs to train the US healthcare workforce have been enacted, they remain limited [34]. Advancing these initiatives will allow physicians to better assess how the environment impacts their patients’ health and how the care they provide impacts the environment.
The data from this survey reinforce that there should be a debate over if, the extent to which, and how any treatment option limitations and climate-informed decisions might occur during clinical care. Limiting environmentally harmful treatment options could occur upstream from or within clinical encounters. No ethical systems promoting climate-informed healthcare support the withdrawal of appropriate care; each chiefly focuses on how to promote efficiency and stewardship of resources [10, 35, 36]. As regulatory decisions in the US are unlikely to remove or diminish treatment availability due to environmental impact, health systems and clinicians become the de facto implementors. A combination of care delivery decisions about formulary, telehealth use, climate disaster preparedness, and “climate-informed consent” are likely avenues for such stewardship.
The latter approach of climate-informed consent appears consistent with the perspectives elicited in this survey. A significant percentage of patients reported they would trial an antihypertensive medicine of lower efficacy if it were better for the environment. Only 23% would not allow any efficacy reduction for a less environmentally toxic treatment, while 44% agreed to a 1–10% reduction in efficacy, and 33% a > 10% reduction. While acceptable reductions may be less for treatments where the stakes are higher (e.g., a potentially curative cancer treatment), these data argue that modern, patient-centered care should include consideration of environmental effects within discussions of treatment risks, benefits, and alternatives [37].
This study has limitations including those inherent to cross-sectional survey design and the use of hypothetical questions. While we purposively sampled to enhance some aspects of who was represented, we did not explicitly assess the distribution of respondents’ political views or other potential unmeasured confounders. Patient respondents were also limited to adults receiving care in academic health systems in the northeast and midwest, which underrepresented rural populations and those from the western and southern US. Although instrument questions were developed using rigorous design methods and cognitive testing, they were not psychometrically validated, although the necessity of such validation for assessments that do not involve psychological or behavioral constructs is debated [38]. Finally, while climate change and environmental impacts are not synonymous concepts, both terms were used based on cognitive testing feedback; however, this overlap in terminology may have caused confusion.
Conclusions
The links between climate, health, and care delivery reveal ethical dilemmas at their intersection. Our data show that acceptance of these links by US patients, like the US physicians we previously surveyed, is broad and growing. Moreover, our findings suggest that there are significant gaps between current knowledge, awareness, and practice patterns that might enable a climate-informed US medical system. Methods for bridging the current and ideal future state include increasing physician training that explicitly outlines individual and health system mechanisms for environmental stewardship and care delivery emissions data that can inform actionable solutions. Closing these gaps will be essential to ensuring that US healthcare can meet its ethical obligations in a warming world.
Supplementary Information
Additional file 1. Supplementary material.
Acknowledgements
None.
Abbreviations
- US
United States
- CVI
Climate Vulnerability Index
- CI
Confidence interval
- IQR
Interquartile ranges
- OR
Odds ratio
Authors’ contributions
Conceptualization: AH, CR, MS, GAA. Methodology: AH, ES, CR, FJH, AR, BNC, GAA, AC. Data Collection: AH, ES, AR, TPW, HJ, FJH, BNC, ASD, EG. Data Analysis: AH, TPW, AC. Manuscript Writing: AH, CR, TPW. Critical Review: ES, FJH, HJ, AC, ASD, EG, AR, BNC, MS, GAA. Manuscript Revision: All authors. Final Approval: All authors read and approved the final manuscript.
Authors’ Twitter handles
X: @ahantel.
BlueSky: @ahantel.bsky.social.
Funding
Funding for the study was provided by a Greenwall Foundation Making a Difference Grant to Hantel, Senay, Hlubocky, and Abel. The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Data availability
The datasets used in the current study are available from the corresponding author on reasonable request, subject to approval by the institutional review boards that approved the study.
Declarations
Ethics approval and consent to participate
The Dana-Farber/Harvard Cancer Center, University of Chicago Medicine, and Mount Sinai Hospital institutional review boards provided ethical approval for the study (reference IDs 22-616 and 22-01009). Participants consented to the survey at the time of participation
Consent for publication
Not applicable.
Competing interests
Dr. Hantel reports receiving personal fees from AbbVie, AstraZeneca, BMS, Celgene, GSK, American Journal of Managed Care, Jazz, and Genentech and employment at Real Chemistry (immediate family member) outside the submitted work. Ms. Cronin reports prior employment at Revitas and employment at Pulse Infoframe (immediate family member) employment and stock at Vertex Pharmaceuticals outside the submitted work. Dr. DuVall reports receiving personal fees from Aptitude Health, BioAscend, CME Concepts, Novartis, and Sago Marketing outside the submitted work. The other authors report no conflicts of interest.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Eckelman MJ, Huang K, Lagasse R, Senay E, Dubrow R, Sherman JD. Health care pollution and public health damage in the United States: an update. Health Aff (Millwood). 2020;39(12):2071–9. [DOI] [PubMed] [Google Scholar]
- 2.Fullman N, Yearwood J, Abay SM, Abbafati C, Abd-Allah F, Abdela J, et al. Measuring performance on the Healthcare Access and Quality Index for 195 countries and territories and selected subnational locations: a systematic analysis from the Global Burden of Disease Study 2016. The Lancet. 2018;391(10136):2236–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Andrieu B, Marrauld L, Vidal O, Egnell M, Boyer L, Fond G. Health-care systems’ resource footprints and their access and quality in 49 regions between 1995 and 2015: an input–output analysis. Lancet Planet Health. 2023;7(9):e747-58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Health System Tracker: What do we know about the burden of disease in the U.S.? : Kaiser Family Foundation; 2017 [Available from: https://www.healthsystemtracker.org/chart-collection/know-burden-disease-u-s/.
- 5.MacNeill AJ, Lillywhite R, Brown CJ. The impact of surgery on global climate: a carbon footprinting study of operating theatres in three health systems. Lancet Planet Health. 2017;1(9):e381–8. [DOI] [PubMed] [Google Scholar]
- 6.Dacones I, Cave C, Furie GL, Ogden CA, Slutzman JE. Patient transport greenhouse gas emissions from outpatient care at an integrated health care system in the Northwestern United States, 2015–2020. The Journal of Climate Change and Health. 2021. 10.1016/j.joclim.2021.100024. [Google Scholar]
- 7.Wang EY, Zafar JE, Lawrence CM, Gavin LF, Mishra S, Boateng A, et al. Environmental emissions reduction of a preoperative evaluation center utilizing telehealth screening and standardized preoperative testing guidelines. Resour Conserv Recycl. 2021. 10.1016/j.resconrec.2021.105652.33821099 [Google Scholar]
- 8.Patel SD, Smith-Steinert R. Greening the operating room, one procedure at a time. The Journal of Climate Change and Health. 2021. 10.1016/j.joclim.2021.100014. [Google Scholar]
- 9.Janson C, Henderson R, Löfdahl M, Hedberg M, Sharma R, Wilkinson AJK. Carbon footprint impact of the choice of inhalers for asthma and COPD. Thorax. 2020;75(1):82–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Hantel A, Marron JM, Abel GA. Establishing and defining an approach to climate conscious clinical medical ethics. Am J Bioeth. 2024. 10.1080/15265161.2024.2337418. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hantel A, Senay E, Hlubocky F, Walsh TP, Johnston H, Cronin A, et al. The ethics of climate change and health-care delivery: a national survey of US-based physicians. Lancet Planet Health. 2025;9(8):101289. 10.1016/j.lanplh.2025.101289. Epub 2025 Aug 11. [DOI] [PubMed]
- 12.Amberger O, Lemke D, Christ A, Muller H, Schwappach D, Geraedts M, et al. Patient safety and climate change: findings from a cross-sectional survey in Germany. BMC Public Health. 2024;24(1):3233. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Cohen ES, Kringos DS, Kouwenberg L, Sperna Weiland NH, Richie C, Aarts JWM, et al. Patient perspectives on climate friendly healthcare: an exploratory study in obstetrics and gynaecology. Patient Educ Couns. 2024;130:108427. [DOI] [PubMed] [Google Scholar]
- 14.Krippl N, Mezger NCS, Danquah I, Nieder J, Griesel S, Schildmann J, et al. Climate-sensitive health counselling in Germany: a cross-sectional study about previous participation and preferences in the general public. BMC Public Health. 2024;24(1):1519. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Leiserowitz A, Maibach E, Rosenthal S, Kotcher J, Goddard E, Carman J, et al. Climate change in the American mind: Beliefs & Attitudes, Spring 2025. New Haven, CT: Yale University and George Mason University; 2025. [Google Scholar]
- 16.Martins Y, Lederman RI, Lowenstein CL, Joffe S, Neville BA, Hastings BT, et al. Increasing response rates from physicians in oncology research: a structured literature review and data from a recent physician survey. Br J Cancer. 2012;106(6):1021–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Hantel A, Senay E, Richie C, Revette A, Nava-Coulter B, Hlubocky FJ, et al. A focus group study of ethical issues during climate-informed health decision-making. Nat Clim Chang. 2024;14(10):1040–6. [Google Scholar]
- 18.The American Community Survey Census.gov: The United States Census Bureau; 2020 [Available from: https://www2.census.gov/programs-surveys/acs/methodology/questionnaires/2020/quest20.pdf.
- 19.Sarfaty M, Bloodhart B, Ewart G, Thurston GD, Balmes JR, Guidotti TL, et al. American Thoracic Society member survey on climate change and health. Ann Am Thorac Soc. 2015;12(2):274–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Sarfaty M, Kreslake JM, Casale TB, Maibach EW. Views of AAAAI members on climate change and health. J Allergy Clin Immunol Pract. 2016;4(2):333-5 e26. [DOI] [PubMed] [Google Scholar]
- 21.Albrecht K, Archibold E. Inductive survey research. In: The SAGE Handbook of Survey Development and Application. 2023. p. 93–108.
- 22.Tee Lewis PG, Chiu WA, Nasser E, Proville J, Barone A, Danforth C, et al. Characterizing vulnerabilities to climate change across the United States. Environ Int. 2023. 10.1016/j.envint.2023.107772. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Kotcher J, Maibach E, Miller J, Campbell E, Alqodmani L, Maiero M, et al. Views of health professionals on climate change and health: a multinational survey study. Lancet Planet Health. 2021;5(5):e316–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Maibach EW, Kreslake JM, Roser-Renouf C, Rosenthal S, Feinberg G, Leiserowitz AA. Do Americans Understand That Global Warming Is Harmful to Human Health? Evidence From a National Survey. Ann Glob Health. 2015;81(3):396–409. 10.1016/j.aogh.2015.08.010. [DOI] [PubMed]
- 25.Ayers M, Marlon JR, Ballew MT, Maibach EW, Rosenthal SA, Roser-Renouf C, et al. Changes in global warming’s six Americas: an analysis of repeat respondents. Clim Change. 2024. 10.1007/s10584-024-03754-x. [Google Scholar]
- 26.Romanello M, Walawender M, Hsu S-C, Moskeland A, Palmeiro-Silva Y, Scamman D, et al. The 2024 report of the Lancet Countdown on health and climate change: facing record-breaking threats from delayed action. The Lancet. 2024;404(10465):1847–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Woodcock A, Janson C, Rees J, Frith L, Lofdahl M, Moore A, et al. Effects of switching from a metered dose inhaler to a dry powder inhaler on climate emissions and asthma control: post-hoc analysis. Thorax. 2022;77(12):1187–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Bonnesen B, Eklöf J, Biering-Sørensen T, Modin D, Miravitlles M, Mathioudakis AG, et al. Effect of low climate impact vs. high climate impact inhalers for patients with asthma and COPD-a nationwide cohort analysis. Respir Res. 2024. 10.1186/s12931-024-02942-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Liu J, Varghese BM, Hansen A, Zhang Y, Driscoll T, Morgan G, et al. Heat exposure and cardiovascular health outcomes: a systematic review and meta-analysis. Lancet Planet Health. 2022;6(6):e484–95. [DOI] [PubMed] [Google Scholar]
- 30.Mason H, C King J, E Peden A, C Franklin R. Systematic review of the impact of heatwaves on health service demand in Australia. BMC Health Services Res. 2022;22(1):960. 10.1186/s12913-022-08341-3. PMID: 35902847. [DOI] [PMC free article] [PubMed]
- 31.Zhang D, Xi Y, Boffa DJ, Liu Y, Nogueira LM. Association of wildfire exposure while recovering from lung cancer surgery with overall survival. JAMA Oncol. 2023. 10.1001/jamaoncol.2023.2144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Zhong A, Amat MJ, Anderson TS, Shafiq U, Sternberg SB, Salant T, et al. Completion of recommended tests and referrals in telehealth vs in-person visits. JAMA Netw Open. 2023. 10.1001/jamanetworkopen.2023.43417. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Ghosh AK, Azan A, Basu G, Bernstein J, Gillespie E, Gordon LB, et al. Building climate change into medical education: a society of general internal medicine position statement. J Gen Intern Med. 2024. 10.1007/s11606-024-08690-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Baker N, Hugh S, Kline MC, Malits J, Mandalapu A, Mazumder DR, et al. Evaluating the impact of a longitudinal, integrated climate change, health, and environment curriculum in undergraduate medical training at Harvard Medical School. PLOS Climate. 2025. 10.1371/journal.pclm.0000727.
- 35.Armand W, Padget M, Pinsky E, Wasfy JH, Slutzman JE, Duhaime A-C. Clinician knowledge and attitudes about climate change and health after a quality incentive program. JAMA Netw Open. 2024. 10.1001/jamanetworkopen.2024.26790. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Richie C. A brief history of environmental bioethics. Virtual Mentor. 2014;16(9):749–52. [DOI] [PubMed] [Google Scholar]
- 37.Richie C. Principles of green bioethics: sustainability in health care. East Lansing: Michigan State University Press; 2019. [Google Scholar]
- 38.Richie C. Green informed consent” in the classroom, clinic, and consultation room. Med Health Care Philos. 2023;26(4):507–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Additional file 1. Supplementary material.
Data Availability Statement
The datasets used in the current study are available from the corresponding author on reasonable request, subject to approval by the institutional review boards that approved the study.