0

The full content of Annals is available to subscribers

Subscribe/Learn More  >
Original Research |

Outcomes of Basic Versus Advanced Life Support for Out-of-Hospital Medical EmergenciesOutcomes of Basic Versus Advanced Life Support

Prachi Sanghavi, PhD; Anupam B. Jena, MD, PhD; Joseph P. Newhouse, PhD; and Alan M. Zaslavsky, PhD
[+] Article, Author, and Disclosure Information

This article was published online first at www.annals.org on 13 October 2015.


From University of Chicago, Chicago, Illinois, and Harvard Medical School, Boston, Massachusetts.

Note: Dr. Sanghavi had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. The data in this article were obtained from the Centers for Medicare & Medicaid Services (CMS) within the U.S. Department of Health and Human Services. The authors' data use agreement with CMS does not allow sharing of individual records. These data can be obtained by others through CMS. However, if there are quantities of interest relevant to the paper that are not in the article, the authors can share results at a higher level of aggregation as long as the CMS data-sharing policies are met.

Grant Support: By a National Science Foundation Graduate Research Fellowship (Dr. Sanghavi), an Agency for Healthcare Research and Quality grant (1R36HS022798-01; Dr. Sanghavi), and the National Institutes of Health Early Independence Award (1DP5OD017897-01; Dr. Jena).

Disclosures: Dr. Jena receives personal fees as a principal consultant to Precision Health Economics. Dr. Newhouse is a director of, and holds equity in, Aetna. Authors not named here have disclosed no conflicts of interest. Disclosures can also be viewed at www.acponline.org/authors/icmje/Conflict OfInterestForms.do?msNum=M15-0557.

Editors' Disclosures: Christine Laine, MD, MPH, Editor in Chief, reports that she has no financial relationships or interests to disclose. Darren B. Taichman, MD, PhD, Executive Deputy Editor, reports that he has no financial relationships or interests to disclose. Cynthia D. Mulrow, MD, MSc, Senior Deputy Editor, reports that she has no relationships or interests to disclose. Deborah Cotton, MD, MPH, Deputy Editor, reports that she has no financial relationships or interest to disclose. Jaya K. Rao, MD, MHS, Deputy Editor, reports that she has stock holdings/options in Eli Lilly and Pfizer. Sankey V. Williams, MD, Deputy Editor, reports that he has no financial relationships or interests to disclose. Catharine B. Stack, PhD, MS, Deputy Editor for Statistics, reports that she has stock holdings in Pfizer.

Reproducible Research Statement:Study protocol: Not applicable. Statistical code: Available from Dr. Sanghavi (e-mail, psanghavi@health.bsd.uchicago.edu). Data set: Available from www.resdac.org.

Requests for Single Reprints: Prachi Sanghavi, PhD, Department of Public Health Sciences, The University of Chicago, 5841 South Maryland Avenue, MC2000, Chicago, IL 60637; e-mail, psanghavi@health.bsd.uchicago.edu.

Current Author Addresses: Dr. Sanghavi: Department of Public Health Sciences, The University of Chicago, 5841 South Maryland Avenue, MC2000, Chicago, IL 60637.

Drs. Jena, Newhouse, and Zaslavsky: Department of Health Care Policy, Harvard Medical School, 180A Longwood Avenue, Boston, MA 02115.

Author Contributions: Conception and design: P. Sanghavi, A.B. Jena, A.M. Zaslavsky.

Analysis and interpretation of the data: P. Sanghavi, A.B. Jena, J.P. Newhouse, A.M. Zaslavsky.

Drafting of the article: P. Sanghavi, A.B. Jena, A.M. Zaslavsky.

Critical revision of the article for important intellectual content: P. Sanghavi, A.B. Jena, J.P. Newhouse, A.M. Zaslavsky.

Final approval of the article: P. Sanghavi, A.B. Jena, J.P. Newhouse, A.M. Zaslavsky.

Statistical expertise: P. Sanghavi, A.M. Zaslavsky.

Obtaining of funding: P. Sanghavi.

Collection and assembly of data: P. Sanghavi.


Ann Intern Med. 2015;163(9):681-690. doi:10.7326/M15-0557
Text Size: A A A

Background: Most Medicare patients seeking emergency medical transport are treated by ambulance providers trained in advanced life support (ALS). Evidence supporting the superiority of ALS over basic life support (BLS) is limited, but some studies suggest ALS may harm patients.

Objective: To compare outcomes after ALS and BLS in out-of-hospital medical emergencies.

Design: Observational study with adjustment for propensity score weights and instrumental variable analyses based on county-level variations in ALS use.

Setting: Traditional Medicare.

Patients: 20% random sample of Medicare beneficiaries from nonrural counties between 2006 and 2011 with major trauma, stroke, acute myocardial infarction (AMI), or respiratory failure.

Measurements: Neurologic functioning and survival to 30 days, 90 days, 1 year, and 2 years.

Results: Except in cases of AMI, patients showed superior unadjusted outcomes with BLS despite being older and having more comorbidities. In propensity score analyses, survival to 90 days among patients with trauma, stroke, and respiratory failure was higher with BLS than ALS (6.1 percentage points [95% CI, 5.4 to 6.8 percentage points] for trauma; 7.0 percentage points [CI, 6.2 to 7.7 percentage points] for stroke; and 3.7 percentage points [CI, 2.5 to 4.8 percentage points] for respiratory failure). Patients with AMI did not exhibit differences in survival at 30 days but had better survival at 90 days with ALS (1.0 percentage point [CI, 0.1 to 1.9 percentage points]). Neurologic functioning favored BLS for all diagnoses. Results from instrumental variable analyses were broadly consistent with propensity score analyses for trauma and stroke, showed no survival differences between BLS and ALS for respiratory failure, and showed better survival at all time points with BLS than ALS for patients with AMI.

Limitation: Only Medicare beneficiaries from nonrural counties were studied.

Conclusion: Advanced life support is associated with substantially higher mortality for several acute medical emergencies than BLS.

Primary Funding Source: National Science Foundation, Agency for Healthcare Research and Quality, and National Institutes of Health.

Figures

Grahic Jump Location
Figure 1.

Country-level ALS penetration rates for major trauma.

These rates are for a standardized population with major trauma but are not derived from characteristics of trauma patients. Rather, they are predicted from rates of ALS use in other diagnosis groups in each county. ALS = advanced life support.

Grahic Jump Location
Grahic Jump Location
Figure 2.

Kaplan–Meier analysis of survival after emergency event, by ambulance service level.

The insets show the survival probability over the full observational period, and the main graphs show it for the first 90 d. Data include emergency medical events between 1 January 2006 and 2 October 2011. Mortality was monitored until 31 December 2011, when the data were censored, and thus there was follow-up to at least 90 days for each beneficiary. Plots use different y-axis scales. ALS = advanced life support; BLS = basic life support. A. Trauma. B. Stroke. C. Acute myocardial infarction. D. Respiratory failure.

Grahic Jump Location

Tables

References

Letters

NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Comments

Submit a Comment/Letter
More Rigorous Study of the Value of Advanced and Basic Life Support Ambulances is Needed
Posted on October 28, 2015
Stephen B. Soumerai, ScD
Harvard Medical School
Conflict of Interest: None Declared

The recent paper by Sanghavi and colleagues reports interesting Medicare claims data on the reduced survival of patients receiving Advance Life Support (ALS) vs. Basic Life Support (BLS) ambulances. Not surprising, as with all research, such studies raise important questions about bias and confounding.

First, the study uses a cross-sectional design with no estimate of pre-post change and no adequate counterfactual (control). Although conducting a randomized control trial (RCT) is difficult, there is a hierarchy of quasi-experimental (non-RCT) designs that produce differing levels of validity of causal inferences.1 The cross-sectional design, with or without propensity score adjustments or instrumental variables (IVs), is not included as evidence in international systematic reviews.2 While it’s hard to find natural experiments of changing ambulance services, this weak design casts doubt on their near-causal inferences that “we found similar or better outcomes associated with prehospital BLS than ALS…”

Second, although the authors attempt to simulate randomization through IVs, one of this paper’s authors, Dr. Zaslavsky, coauthored our 2014 Annals study that found 20 years of similar regional IVs in comparative effectiveness research were likely confounded by unmeasured variables.3 It is unlikely that county-level variation in ALS vs. BLS is random. 

Third, Table 1 provides almost no clinical description of patients served by ALS and BLS.  To be more useful, the table must include direct comparisons of the severity of the four conditions prompting the call.  Is a near-fatal trauma more likely to receive ALS?  Many programs triage ALS, falsely associating ALS with excess death rates. Their year-old comorbidity measure has limited relevance to the high acuity survival outcome. In addition, ALS patients travel farther for all four study conditions; live in less metropolitan settings; have fewer options for trauma centers and teaching hospitals; and are younger. This suggests the presence of other differences that were not measured.

Given controversies regarding the value of ALS and its importance for health policy, more rigorous controlled studies are warranted. This study is a useful contribution but also illustrates the difficulty of studying America’s highly varied and complex health systems.  We must apply the strongest research design whenever possible.

1. Soumerai SB, Starr D, Majumdar S. How do you know which health care effectiveness research you can trust?: A guide on study design for the perplexed. Prev Chronic Dis. 2015;12:E101.http://www.cdc.gov/pcd/issues/2015/15_0187.htm
2. Effective Practice and Organisation of Care (EPOC). What study designs should be included in an EPOC review? EPOC Resources for review authors. Oslo: Norwegian Knowledge Centre for the Health Services; 2013. Available at: http://epoc.cochrane.org/epoc-specific-resources-review-authors
3. Garabedian LF, Chu P, Toh S, Zaslavsky AM, Soumerai SB. Potential bias of instrumental variable analyses for observational comparative effectiveness research. Ann Intern Med. 2014 Jul 15;161(2):131-8.

Two questions about the comparison in an observational study for basic versus advanced life support.
Posted on November 30, 2015
Yusuke Tsutsumi, Yasushi Tsujimoto, Yuki Kataoka, Tatsuyoshi Ikenoue
Department of Healthcare Epidemiology, Graduate School of Medicine and Public Health, Kyoto University
Conflict of Interest: None Declared
To the editor: We read the article by Dr. Sanghavi et al. with great interest and appreciate author’s efforts to conduct a comprehensive research (1). However, we would like to point out two questions about analysis methods.
Firstly, the authors used propensity score weighting methods. According to the previous study the authors cited as reference (2), if the propensity score for individual i is πi (that is the probability to receive advanced life support (ALS) in this case), the strategy of weighting is inverse of πi for ALS and inverse of 1-πi for basic life support (BLS). This weighting method is well known as an inverse probability of weighting method. However, the author’s method of weighting mentioned in Supplement material was 1-πi for ALS and πi for BLS reversely. We wonder if this weighting method by authors was appropriate.
Secondly, the authors used a unique instrumental variable method for main results where they first estimated a county-level probability of ALS use from a nonlinear regression model and then use this predicted probability as instruments in the first stage of two-stage-least-squares (2SLS) methods. We want to know the result when using usual 2SLS methods. Conventional 2SLS tend to be biased when the instrument is very weak (3). Please show us the merits of the authors’ unique method and how much this method improve the weakness of instrument.
References
1. Sanghavi P, Jena AB, Newhouse JP, Zaslavsky AM. Outcomes of basic versus advanced life support for out-of-hospital medical emergencies. Ann Intern Med. 2015; 163: 681-690
2. Hirano K, Imbens GW. Estimation of causal effects using propensity score weighting: An application to data on right heart catheterization. Health Serv Outcomes Res Methodol 2001; 2: 259-278
3. Stock JH, Watson MW. Introduction to Econometrics. Update 3rd ed. Pearson Education Limited; 2014. 852 p.
Author's Response
Posted on January 29, 2016
Prachi Sanghavi, Anupam B. Jena, Joseph P. Newhouse, Alan M. Zaslavsky
The University of Chicago and Harvard University
Conflict of Interest: Grant support was provided by a National Science Foundation Graduate Research Fellowship (Dr. Sanghavi), an Agency for Healthcare Research and Quality grant (1R36HS022798-01; Dr. Sanghavi), and the National Institutes of Health Early Independence Award (1DP5OD017897-01; Dr. Jena).
Although we agree with Soumerai that randomized trials of prehospital care would be useful, they are unlikely. In their absence one must use the best available observational study designs. We disagree with Soumerai’s characterization of all instrumental variables (IV) designs as weak. Just as applications of the difference-in-difference designs that Soumerai favors vary in strength, so too do IV applications.
The key untestable assumptions for valid IV analyses are that the instrument affects the dependent variable only through the variable being instrumented for and is uncorrelated with omitted predictors of the outcome [1,2]. These assumptions are likely satisfied in our study. Although the variation across counties in the proportion of ALS ambulances is not randomized, it does appear to be idiosyncratic, driven in part by local political decisions. Thus, our IV analysis approximates a randomized trial with multiple arms that vary the proportion of ALS ambulances. One must assume that neither the severity (within diagnosis) of emergent conditions across counties (within states) leading to ambulance transport nor the quality of hospital care is associated with ALS penetration. The former assumption is plausible but untestable for four of the diagnoses; we control for severity in the case of trauma. The latter assumption is testable and is not rejected.
Soumerai cites a paper critical of instrumental variables commonly used in comparative effectiveness research, of which he and Zaslavsky were coauthors [3]. That paper noted that effects of regional variation in an instrument might be mediated through regional correlates such as income or urbanicity, violating IV’s assumptions. Our analysis controlled for these and several other regional characteristics. It also included a falsification test that showed acute inpatient mortality not following emergency transport did not exhibit the same regional patterns as post-ambulance mortality. That paper concluded, “The assumption that the instrument is only related to the outcome through the treatment may apply best to specific, focused, plausibly exogenous interventions or events, such as natural experiments … or changes in policy or technology.” We consider it probable that after the controls in our modeling strategy, our instrument fits this description.
We appreciate Doctors Tsutsumi, Tsujimoto, Kataoka, and Ikenoue’s attention to technical detail. Inverse probability (Horvitz-Thompson) propensity score weighting is only one of a family of weighting schemes that simulate covariate balance for causal comparisons, but it does not minimize variance or even always have a finite expected value.[4] Our (1-πi, πi) weighting yields exact mean balance when used with propensity scores from a logistic regression, and minimizes variance over all balancing weights.[4] It estimates an average treatment effect weighted to a population distribution constituting the overlap of the distributions in treated and control groups, nearest to equipoise between treatments.
Appendix A8 reports first-stage F statistics exceeding 1,000 in each diagnosis group, implying our instrument is strong; eTable5 shows the variation in the predicted probability of ALS for each diagnosis.
[1] Angrist JD, Imbens G, Rubin DB. Identification of causal effects using instrumental variables, JASA 1996; 91(434):444-455.

[2] Angrist JD, Pischke J-S. Mostly Harmless Econometrics, Princeton NJ, Princeton University Press, 2008.

[3] Garabedian LF, Chu P, Toh S, Zaslavsky AM, Soumerai SB. Potential bias of instrumental variable analyses for observational comparative effectiveness research. Ann Intern Med, 2014;161(2):131-8.
[4] Li F, Morgan KL, Zaslavsky AM. Balancing Covariates via Propensity Score Weighting. arXiv:1404.1785v2 [stat.ME] (http://arxiv.org/abs/1404.1785v2).

Submit a Comment/Letter

Summary for Patients

Clinical Slide Sets

Terms of Use

The In the Clinic® slide sets are owned and copyrighted by the American College of Physicians (ACP). All text, graphics, trademarks, and other intellectual property incorporated into the slide sets remain the sole and exclusive property of the ACP. The slide sets may be used only by the person who downloads or purchases them and only for the purpose of presenting them during not-for-profit educational activities. Users may incorporate the entire slide set or selected individual slides into their own teaching presentations but may not alter the content of the slides in any way or remove the ACP copyright notice. Users may make print copies for use as hand-outs for the audience the user is personally addressing but may not otherwise reproduce or distribute the slides by any means or media, including but not limited to sending them as e-mail attachments, posting them on Internet or Intranet sites, publishing them in meeting proceedings, or making them available for sale or distribution in any unauthorized form, without the express written permission of the ACP. Unauthorized use of the In the Clinic slide sets will constitute copyright infringement.

Toolkit

Buy Now for $32.00

to gain full access to the content and tools.

Want to Subscribe?

Learn more about subscription options

Advertisement
Related Articles
Related Point of Care
Topic Collections
PubMed Articles
Forgot your password?
Enter your username and email address. We'll send you a reminder to the email address on record.
(Required)
(Required)