Presenting quality data to vulnerable groups: charts,summaries or behavioral economic nudges?

Abstract

Objectives

Despite the increased focus on health care consumers’ active choice, not enough is known about how to best facilitate the choice process. We sought to assess methods of improving this process for vulnerable consumers in the United States by testing alternatives that emphasize insights from behavioral economics, or ‘nudges’.

Methods

We performed a hypothetical choice experiment where subjects were randomized to one of five experimental conditions and asked to choose a health center (location where they would receive all their care). The conditions presented the same information about health centers in different ways, including graphically as a chart, via written summary and using behavioral economics, ‘nudging’ consumers toward particular choices. We hypothesized that these ‘nudges’ might help simplify the choice process. Our primary outcomes focused on the health center chosen and whether consumers were willing to accept ‘nudges’.

Results

We found that consumer choice was influenced by the method of presentation and the majority of consumers accepted the health center they were ‘nudged’ towards.

Conclusions

Consumers were accepting of choices grounded in insights from behavioral economics and further consideration should be given to their role in patient choice.

Keywords

behavioral economics consumer choice health services quality data

Introduction

Health care quality information is currently used for multiple purposes. It is used by policy makers and researchers to evaluate new pilot programs of Medicare and Medicaid,^1,2 by health plans and other organizations to grade provider performance related to specific health outcomes^3,4 and by payers to deny reimbursements for under-performance and iatrogenic errors.^5–7 In addition, health care quality data are increasingly being offered to consumers in an effort to influence their choices about providers and health plans.^8–10 Indeed, quality data will be one tool by which consumers can compare coverage options via Health Insurance Exchanges as part of the United States Affordable Care Act.^11,12

A reliance on quality data makes sense if the right metrics are utilized: high performers should be rewarded for keeping individuals and populations healthy, and under-performers penalized. Competition to improve quality (and outcomes by proxy) will increase, and consumers can then use such data to make informed choices about their health care providers. Yet some research shows that offering choice is not always unambiguously welfare-enhancing.¹³ When faced with information meant to guide choice, some consumers opt not to make any choice at all. This type of choice deferral can have negative consequences, especially in the health care or health coverage context.¹⁴ In addition, emerging research on US health insurance exchanges has demonstrated that the choice process has the potential to be confusing and burdensome for consumers.¹⁵

Vulnerable populations for the purposes of this article are defined as low socioeconomic status people who access the US health care safety net, including public hospitals. They are particularly understudied when it comes to health care choice. Under the Affordable Care Act, large numbers of vulnerable people are being offered insurance cover. The degree to which this group accesses and uses quality data to inform health care decision-making remains a critical policy question. Compared to non-vulnerable populations, health care choices for the vulnerable may be limited due to more restrictive option pools or barriers to active choice engagement, including low health and computer literacy, or challenges accessing available quality information.^16–18 Despite factors such as these that have the potential to limit choice, prior work in the United States shows that individuals, including those in vulnerable populations, do want to access health care quality and performance data.^19–21

On 7 September 2007, the New York City Health and Hospitals Corporation (HHC) released quality and safety data for each of its facilities and, in doing so, became the first public hospital system to release these data publicly.²² The goal of releasing the data was to increase transparency and to influence consumer choices. HHC is the city’s health system for vulnerable safety net patients and serves over 1.4 million New York City residents. The HHC quality data consist of established process and outcome quality measures drawn largely from federal government quality data standards. Examples include post-admission mortality rates, infection rates, prevention and screening rates, and adherence to best practice treatments for conditions such as heart attack. Along with data from each of the 11 HHC acute care hospitals, HHC has posted city, state, and/or national comparison data when available. Presentation formats include text and graphics (e.g. bar and pie charts, line charts) and are available online.²³ We found in prior qualitative work on this topic with HHC consumers that they were interested in utilizing this type of information in their choices, but were not aware it existed.²⁴

Despite its potential impact on choice, research demonstrates that use of quality information by consumers is minimal at best.^25–27 Deferred choice in health care, especially when it concerns health plan enrollment, remains problematic in the US market that requires a broad risk pool to function optimally. The complexity of the health care delivery system can be misunderstood by consumers, which can also serve to inhibit the uptake of quality data.²⁸ Prior research has focused on improving the content and comprehensibility of quality data so that typical consumers can make educated choices about their health care based on accurate, reliable data they care about.^19,29 While some consumers may be influenced to incorporate quality data in decision-making if it is presented in a more appealing manner, others might have trouble interpreting or understanding such data, or may not be interested in these data. In addition, relatively subtle differences, such as the order in which choices are presented, can influence decision-making.^29,30 In such cases, it may be possible to increase the odds that quality data will be incorporated into consumer choice by using a ‘nudge’.³¹

Nudges, in this context, are means of structuring how a choice is presented, also known as altering the choice architecture. The concept, very newly applied in the health sector, emerged from behavioral economics, largely popularized by Thaler and Sunstein in their work by the same name.³¹ The basic concept is that the choice architecture can influence ultimate decisions.³¹ Nudges have been shown in other contexts to help move consumers towards options that might be most appropriate for them given their needs and preferences while lessening the burden that can sometimes come with complex choice. Such methods have, to date, not been widely applied in the health field and in health care choice in particular, and the effect of nudges in health care settings is unclear.

Our study was focused specifically on a particular type of nudge, called a ‘default’ option. A default option is an option or set of options that the consumer is steered towards from a larger choice set. Default options have been enormously successful in the United States in encouraging employees to sign up for company-provided retirement plans, which provide significant financial advantage to most consumers. When employees are not enrolled in any plan and are required to actively choose a plan, over 50% do not enroll. Conversely, when the default option is to be enrolled in a plan without any additional action on their part, almost 90% of employees are enrolled even though the process not to enroll is very simple. While the nudge that we test is more subtle than the example above, it is a first step in exploring the role of nudges in health care decision-making for vulnerable consumers.

The purpose of this exploratory study, following from our earlier qualitative work, is to compare how different types of quality data, methods of presenting such data and the use of behavioral economics-based approaches can influence choice about where to receive health care for vulnerable populations. More specifically, participants in this experiment were asked to choose a health center or a single location where all their care would be received, reflecting the health care environment of the study setting (explained in more detail below). The primary goals were to determine the relative influence of various data presentation styles, including nudges, on participants’ choices (whether they shifted the distribution of choices), their overall satisfaction and whether or not they were accepting of nudges. We performed this hypothetical choice experiment on a set of patients and visitors to the outpatient unit of Bellevue Hospital Center, the largest public hospital in New York City. A key objective was to provide some initial data regarding vulnerable consumers’ potential willingness to accept nudges in health care decision-making.

Methods

Experimental design

We performed a randomized, survey-based, hypothetical choice experiment in which patients were randomly assigned to five conditions presenting health center quality information. Subjects were provided with written text asking them to imagine they had just moved to a new city and needed to choose a place to obtain all their health care (inpatient and outpatient).

We chose to focus on the idea of a health center as it most reflected the real choice faced by subjects at the study hospital and many others served in similar safety net systems. Focusing on insurance choices in new health exchanges or otherwise, or the choice of a single provider, would not have characterized how subjects in our study actually make these choices; the choice would have felt unrealistic to them. In our earlier qualitative work with the same set of consumers, we found that many conceptualized the location of where they chose to receive care as the single location where all of their comprehensive care needs were met–both inpatient and outpatient.

Participants completed the survey by themselves, but study staff was available to answer questions and assist if needed. The text informed participants that all the information available that could aid their choice of health center was provided, and they had no friends, family or others from whom to get other relevant information to assist with the choice. They then turned the page to find the set of information that they had been randomized to see (one of five possible conditions outlined below), reflective of the health center options that were hypothetically available to them. After spending as much time as they wanted, subjects were asked to choose one of four health centers, or to indicate if they could not choose. This was followed with a set of additional questions, including demographics. Participants then returned the survey to the staff and were compensated with US$5. This study was approved by the Institutional Review Boards of New York University School of Medicine and Bellevue Hospital Center.

Sample

The sample was drawn from the clinical area waiting rooms of Bellevue Hospital Center (BHC). Bellevue is the oldest public hospital in the nation, and serves over 500,000 outpatients annually – most of whom are uninsured or are Medicaid beneficiaries (i.e. on a low income).³² Subjects could approach study staff sitting by a sign indicating a research survey could be completed for compensation, or were approached by study staff in the outpatient clinic waiting area; as such, response rates are difficult to determine and we consider this a convenience sample. Requirements to participate in the study were a self-reported ability to read and write in English, and being aged 18 or over.

Presentation of health center information

We presented a small yet diverse set of quality measures to the subjects. We chose to present only 4–6 measures (depending on the randomized condition) to subjects so as not to overwhelm the choice with too much information. We attempted to find measures that resonated broadly with a wide subset of patients, as measured through our focus groups and pretesting²¹ and other published literature.³³ Health center measures for each condition were based on standardized quality measures drawn from the US Center for Medicare and Medicaid Services (CMS) measures.

All subjects, regardless of the condition to which they were randomized, were given the following accepted quality data measures to aid their health center choice: (a) illness prevention (percentage of patients aged 50–80 years who received appropriate screening for colon cancer); (b) acute illness (percentage of heart attack patients receiving appropriate care); (c) chronic illness (percentage of diabetes patients receiving appropriate care); (d) death rates (30-day death rates after hospitalization from health attack). In addition, some subjects were also provided data on two measures that consumers had found important in our prior work: customer service (represented by the percentage of patients who reported that the health center was ‘always’ clean), and physician communication (percentage of patients who reported that their doctors ‘always’ communicated well).²¹

It was important to ensure that the hypothetical health center choices were realistic but also not overly simplistic. We structured choices such that there were no dominant options and no health centers that were overly-obvious winners or losers. As such, we cannot a priori determine whether a plan is appropriate for a randomly selected consumer (without greater insight into their preferences). We used the most recent year of national data to determine the means and distributions for each measure. We then randomly selected a value for each health center from this true population distribution (of each measure independently, ignoring the covariance structure) in an effort to approximate actual representative values. After pretesting, we adjusted some values slightly to assure that each of the four health centers was highest in one category and lowest in one category, straying as little as possible from the value that was randomly selected (see Supplementary material online, Appendix 1, which displays values for each health center).

Experimental conditions

A total of five experimental conditions including the control condition were randomly distributed to participants. The five conditions were as follows:

Control: In the control condition, subjects were presented with charts showing only the four non-consumer-reported measures (prevention, acute care, chronic disease, death rates). The charts were meant to be easy to read, and all included a comparison with the city average as well as a one-sentence descriptor in lay terminology (mirroring CMS language). However, they did not interpret the data for the subject. These charts were always available for all of the experimental conditions (see Supplementary material online Appendix 2, which displays examples of charts available to consumers at all times).

Interpretive Summary + Graphs: In this condition, we provided written information that pointed to the key tradeoffs of each available health center represented in the charts described above. Here, the charts from the control condition were available, but handed to subjects in a closed envelope. The information assisted participants with data interpretation in this option, and highlighted where each health center was the best performer (all were best at one metric), the worst performer (all were worst at one) and for which metrics the health center was average. For example, the summary condition might read ‘Health Center A is the best at managing medical conditions, similar to other health centers for emergency care, about average in preventing illness, and has the highest death rate after a heart attack.’

Interpretive Summary + Consumer-reported Data: In this condition, the written summaries detailed above were augmented to include key data on the two consumer-reported measures that consumers had previously reported as important: health center cleanliness and physician communication. Again, chart depictions were available by opening an envelope.

Behavioral Economics – Individualized: In this condition, subjects answered a series of demographic and preference (e.g. important of various aspects of the choice at hand) questions, and gave the completed information back to the research assistant. Based on the information they submitted, subjects were then ‘nudged’ towards a choice when the research assistants told them which of the health centers was ‘best’ for them. In actuality, the health center was randomly assigned. This condition aimed to test whether consumers might accept these ‘nudge’ approaches (as opposed to whether they would differentially accept a plan that did align with their preferences, were our approach to be implemented in practice). All subjects were given the option to opt out of the health center they were nudged toward and choose any health center. Just as in the Interpretive Summary + Graphs condition, we handed respondents summaries on top of an envelope containing the charts. Subjects were told they had the option (but were not required) to open the sealed envelope which contained the charts with the additional information. Our nudges did not reflect specific consumer preferences per se, but were randomly assigned. As a result, we were not testing whether subjects would accept a well-designed nudge, but the ‘floor’ for accepting nudges.

Behavioral Economics – Location: In this condition, subjects were told which health center was closest to where they lived, and again in reality were randomly assigned a center. The same information in the Behavioral Economics – Individualized condition was made available here. Again, all subjects were given the option to opt out of the health center they were nudged toward and choose any health center.

Outcomes

We focused on (a) whether subjects chose a health center or deferred choice altogether; (b) whether subjects in the two behavioral economics conditions chose the health center toward which they were nudged, or opted for another center; and (c) the actual health center chosen. In addition to the choice measures, we examined a variety of survey measures; (d) satisfaction with the chosen health center; (e) satisfaction with information available; and (e) difficulty with making the choice (all reported on 1–5 Likert scales).

Statistical analysis

First, we checked that our randomization had resulted in balanced samples across each of the experimental conditions by looking for differences in demographics across conditions. Next, to evaluate the differences between the five experimental conditions, we performed a series of ANOVAs for the continuous variables or chi-square tests for the categorical variables, to examine whether the outcomes for the experimental conditions were different from each other. Given the nature of this study, we choose to look at overall differences as opposed to specific differences among individual experimental conditions.

Results

A total of 1010 surveys were completed, approximately 200 in each experimental condition. We found no differences in any of the demographic or other variables across the experimental conditions (Table 1). The average age of the sample was 47. Approximately 36% of the sample was black, 34% Hispanic, 30% white or some other race; 58% was male. About 52% of the sample had a high school education or less. We also asked if subjects had ever been diagnosed by a health care provider with a number of medical conditions. High blood pressure was reported by 40% of subjects, mental illness by 28%, diabetes by 16% and stroke by 13%.

Table 1.

Sample characteristics (N = 1010).

Characteristic	Percent of sample
Age
18–89	47
Race/Ethnicity
Black	35.6
White	29.6
Hispanic	34.1
Asian	6.3
Missing/Other	7.5
Gender
Male	58.0
Female	40.0
Missing/Other	2.2
Education
Eighth grade or less	7.3
Some high school	13.5
High school graduate	31.6
Some college	25.9
Four-year college	11.3
More than college	7.0
Missing	2.4
Health conditions present
High blood pressure	40.0
Stroke	13.0
Mental illness	28.4
Cancer	7.5
Diabetes	16.3
COPD/Breathing problems	24.0
Back problems	23.4
HIV/AIDS	5.4
Alcohol abuse	20.7

We found no differences among the conditions in the percentage of subjects who deferred choice, with percentages ranging from 11 to 14 across all conditions (data not shown). We next examined the choices made by subjects and found overall differences by experimental condition (p < 0.05) (Table 2). In first examining the control condition, subjects reported a slight preference for Health Center A, followed by B. However, in the Interpretive Summary + Graphs condition, which presented the same information as the control condition, but included an interpretation of the charts, the center chosen more often than others was D. Upon the introduction of the consumer-reported measures, the preference returned to A. In the Behavioral Economics – Individualized condition, the health center choice appeared to be randomly distributed consistent with health center random assignment, while in the Behavioral Economics – Location condition the preference shifted back to Center A. This could reflect consumers being less likely to accept nudges based on health center location as opposed to nudges informed by their individual preferences.

Table 2.

Percent of participants choosing each plan.

	Plan
Condition	A	B	C	D	No choice
Control	38	23	10	16	13
Summary	24	16	14	32	14
Summary + Consumer	39	23	12	13	13
Behavioral Economics – Individualized	24	24	20	20	11
Behavioral Economics – Location	34	17	19	17	13

χ² for differences by experimental condition significant at p < 0.05.

While behavioral economics-shaped conditions did not influence consumers deferring choice altogether, 58% of subjects chose the health center that they were nudged toward in the Behavioral Economics – Individualized condition, and 52% in the Behavioral Economics – Location condition (data not shown).

In Table 3 we present differences in the subjects’ satisfaction with their choice, satisfaction with the information provided and overall difficulty. While the first two ANOVAs are statistically significant (p < 0.05), the overall differences are quite small. This implies that subjects were generally equally satisfied with each of the choice processes and information available, and did not find any of the conditions more difficult relative to the others.

Table 3.

Satisfaction with choice and information; difficulty choosing.

	Participant ratings, Scale of 1–5
Condition	Average degree of satisfaction with choice^a	Average satisfaction with information available^a	Average difficulty making a choice
Control	3.76	3.52	2.21
Summaries	3.56	3.25	2.27
Summary + Consumer	3.41	3.19	2.39
Behavioral Economics – Individualized	3.6	3.42	2.24
Behavioral Economics – Location	3.6	3.46	2.2

^aχ² for differences by experimental condition significant at p < 0.05.

Discussion

This study represents an initial examination of an important set of questions that are increasingly relevant for understanding how vulnerable populations make decisions about where to receive health care (our theoretical health centers) based on quality information. We examined various methods of presenting quality data, including testing how subjects respond to behavioral economics ‘nudges’ that might be useful to improve the incorporation of quality data into consumer choice and help prevent choice deferral.

We found that how the information is presented can influence choice in non-trivial ways. Presenting the same information using different presentation methods influenced which health center was chosen, though our study was not designed to test why this might be the case. Other studies have shown that when less information is presented, consumers tend to use the information that is available to a greater degree.^34,35 Indeed, this is the underlying logic behind ‘nudge’-based approaches with vulnerable populations, one could also imagine testing simple graphical presentation, including pictograms.^36,37 It is further notable that the introduction of consumer-reported information, including measures of health center cleanliness and satisfaction with physician communication, influenced health center choice over and above more clinical quality measures. This is consistent with prior work with different patient populations.³³ Overall, participants’ satisfaction and reported ease with the choice process was fairly similar across all the experimental conditions.

The study is not without limitations. The goals were generally of a pilot nature, and the experiment was not set up to understand the mechanism behind particular choices. While our sample was drawn from a population of interest – individuals who use a safety net health system – we were only able to include participants who could read English, and could not examine differences by literacy or numeracy. This may limit the generalizability of our results.

Incorporating behavioral economic-influenced methods that encourage subject choice in a particular direction was successful. Most participants easily accepted our nudge when they believed it was based on which health center was best for them, even when it was randomly assigned. This could be for any number of reasons, including the desire to choose something that was recommended. Most subjects were also willing to accept a separate nudge based purely on health center location. In general, satisfaction with the choice and choice process was unaffected by the method of presenting information or the choice process. Nudges did not result in lower satisfaction or increased choice deferral. The fact that subjects were so easily swayed, even when the choice they were directed towards was not related to their preferences, is noteworthy and points to the potential impact such tools can have on consumers. This implies that methods that do take into account consumer preferences and/or outcomes for specific types of consumers would likely be found acceptable by consumers, and could improve their decision-making. Moreover, it also means that these approaches could be used by private companies or the government to manipulate people, without proper oversight.

As the United States moves even more towards a system based on consumer choice, understanding how various methods of information presentation can influence choice and how to incorporate quality information in choices is of key importance. This is particularly true for behavioral economics-grounded approaches. There is very limited prior information on how patients might respond to approaches such as nudges. We focused on choice within a vulnerable population which includes the uninsured. As health insurance exchanges are initially focused exclusively on these populations, our results are particularly instructive.

In general, consumers are influenced by the presentation of information, and appear open to behavioral economics-influenced approaches that shape consumer choice. However, more work is needed fully to understand the extent to which this is true and, if so, how best to present information so that consumers can make decisions that are optimal for them, and whether other more streamlined methods of nudging consumers towards a particular choice might be welfare-maximizing in the larger sense.

Footnotes

Acknowledgements

The authors thank the peer reviewers for their helpful comments. The authors also thank Bellevue Hospital Center and its administration for their support and cooperation, and Kristin Van Busum, MPA, RAND Corporation, and Rebecca DiBennardo, MA, University of California, Los Angeles, for their support in project coordination and research.

Funding

This study was funded by the Robert Wood Johnson Foundation Quality/Equality initiative.

Supplementary Material

The online appendices are available at .

References

Peikes

Chen

Schore

. Effects of care coordination on hospitalization, quality of care, and health care expenditures among Medicare beneficiaries: 15 randomized trials. JAMA 2009; 301: 603–618.

Centers for Medicare and Medicaid Services. Medicare demonstration projects and evaluation reports. http://www.cms.gov/Medicare/Demonstration-Projects/DemoProjectsEvalRpts/index.html (2012, accessed 6 August 2012).

Davies

Bunn

Fromer

. A focus on the asthma HEDIS measure and its implications for clinical practice. Manag Care Interface 2006; 19: 29–36.

Gelfand

Colice

Fromer

. Use of the Health Plan Employer Data and Information Set for measuring and improving the quality of asthma care. Ann Allergy Asthma Immunol 2006; 97: 298–305.

Lavine

. CMS expands list of hospital mistakes it will not cover. Am J Health Syst Pharm 2008; 65: 1686–1686.

Mattie

Webster

. Centers for Medicare and Medicaid Services’ “Never Events”: an analysis and recommendations to hospitals. Health Care Manag 2008; 27: 338–349.

Welton

. Implications of Medicare reimbursement changes related to inpatient nursing care quality. J Nurs Adm 2008; 38: 325–330.

De Vos

Graafmans

Kooistra

. Using quality indicators to improve hospital care: a review of the literature. J Qual Health Care 2009; 21: 119–129.

Robert Wood Johnson Foundation. RWJF launches website to help patients compare local hospitals and physicians. http://www.rwjf.org/qualityequality/ product.jsp? id =72540 (2011, accessed 6 August 2012).

10.

Hibbard

Greene

Sofaer

. An experiment shows that a well-designed report on costs and quality can help consumers choose high-value health care. Health Aff 2012; 31(3): 560–568.

11.

Centers for Medicare and Medicaid Services. Medicaid Program; Eligibility changes under the Affordable Care Act of 2010. Fed Regist 2012; 77: 17144–17217.

12.

Centers for Medicare and Medicaid Services. Patient Protection and Affordable Care Act; Establishment of exchanges and qualified health plans; Exchange standards for employers. Fed Regist 2012; 77: 18310–18475.

13.

Hibbard

Slovic

Peters

. Is the informed-choice policy approach appropriate for medicare beneficiaries? Health Aff 2001; 20(3): 199–203.

14.

Millett

Chattopadhyay

Bindman

. Unhealthy competition: consequences of health plan choice in California Medicaid. Am J Public Health 2010; 100: 2235–2240.

15.

Sinaiko

Ross-Degnan

Soumerai

. The experience of Massachusetts shows that consumers will need help in navigating insurance exchanges. Health Aff 2013; 32: 78–86.

16.

Phibbs

Mark

Luft

. Choice of hospital for delivery: a comparison of high-risk and low-risk women. Health Serv Res 1993; 28: 201–222.

17.

Greene

Peters

. Medicaid consumers and informed decision making. Health Care Financ Rev 2009; 30: 25–40.

18.

Trivedi

Daly

Kern

. Barriers to implementation of a computerized decision support system for depression: an observational report on lessons learned in real world clinical settings. BMC Med Inform Decis Mak 2009; 9: 6–6.

19.

Edgman-Levitan

Cleary

. What information do consumers want and need? Health Aff 1996; 15: 42–56.

20.

Bloche

. Consumer-directed health care and the disadvantaged. Health Aff 2007; 26: 1315–1327.

21.

Raven

Gillelspie

DiBennardo

. Vulnerable patients’ perceptions of health care quality and quality data. Med Decis Making 2012; 32: 311–326.

22.

New York City Health and Hospitals Corporation. Infection rates continue to drop at city public hospitals. http://www.nyc.gov/html/hhc/html/pressroom/press-release-20080514.shtml (2008, accessed 6 August 2012).

23.

New York City Health and Hospitals Corporation. HHC in focus: understanding our quality and safety performance. http://www.nyc.gov/html/hhc/infocus/ html/home/performancelanding.shtml (2012, accessed 6 August 2012).

24.

Schindler

Kiszko

Abrams

. Environmental and individual factors affecting menu labeling utilization: a qualitative research study. J Acad Nutr Diet 2013; 1113: 667–672.

25.

Robinson

Brodie

. Understanding the quality challenge for health consumers: the Kaiser/AHCPR Survey. Jt Comm J Qual Improv 1997; 23: 239–244.

26.

Shaller

Sofaer

Findlay

. Consumers and quality-driven health care: a call to action. Health Aff 2003; 22: 95–101.

27.

Fung

Lim

Mattke

. Systematic review: the evidence that publishing patient care performance data improves quality of care. Ann Int Med 2008; 148: 111–123.

28.

Lubalin

Harris-Kojetin

. What do consumers want and need to know in making health care choices? Med Care Res Rev 1999; 56(Suppl 1): 67–112.

29.

Hibbard

Peters

Slovic

. Making health care quality reports easier to use. Jt Comm J Qual Improv 2001; 27: 591–604.

30.

Jarvenpaa

. The effects of task demands and graphical format on information-processing strategies. Manag Sci 1989; 35: 285–303.

31.

Thaler

Sunstein

. Improving decisions about health, wealth, and happiness, New Haven: Yale University Press, 2008.

32.

New York City Health and Hospitals Corporation, Bellevue. About Bellevue: facts. http://www.nyc.gov/html/hhc/bellevue/html/about/facts.shtml (2012, accessed 6 August 2012).

33.

Sofaer

Crofton

Goldstein

. What do consumers want to know about the quality of care in hospitals? Health Serv Res 2005; 40: 2018–2036.

34.

Peters

Dieckmann

Dixon

. Less is more in presenting quality information to consumers. Med Care Res Rev 2007; 64: 169–190.

35.

Fasolo

Reutskaja

Dixon

. Helping patient’s choose: how to improve the design of comparative scorecards of hospital quality. Patient Educ Couns 2010; 78: 344–349.

36.

Fortin

Hirota

Bond

. Identifying patient preferences for communicating risk estimates: a descriptive pilot study. BMC Med Inform Decis Mak 2001; 1: 2–2.

37.

Schapira

Nattinger

McHorney

. Frequency or probability? A qualitative study of risk communication formats used in health care. Med Decis Making 2001; 21: 459–467.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.03 MB

0.37 MB

0.03 MB