[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Table 1.  Sociodemographic Features Associated With Presence or Absence of Financial Notes for 46 244 Patients
Sociodemographic Features Associated With Presence or Absence of Financial Notes for 46 244 Patients
Table 2.  Association Between Sociodemographic and Clinical Features and Presence of Financial Notes for 46 244 Patients
Association Between Sociodemographic and Clinical Features and Presence of Financial Notes for 46 244 Patients
Table 3.  Nature of Documented Financial Conversations of 300 Representative Notes
Nature of Documented Financial Conversations of 300 Representative Notes
1.
Hunter  WG, Zhang  CZ, Hesson  A,  et al.  What strategies do physicians and patients discuss to reduce out-of-pocket costs? analysis of cost-saving strategies in 1,755 outpatient clinic visits.  Med Decis Making. 2016;36(7):900-910. doi:10.1177/0272989X15626384PubMedGoogle ScholarCrossref
2.
Schwartz  JAT, Pearson  SD.  Cost consideration in the clinical guidance documents of physician specialty societies in the United States.  JAMA Intern Med. 2013;173(12):1091-1097. doi:10.1001/jamainternmed.2013.817PubMedGoogle ScholarCrossref
3.
Neumann  PJ, Palmer  JA, Nadler  E, Fang  C, Ubel  P.  Cancer therapy costs influence treatment: a national survey of oncologists.  Health Aff (Millwood). 2010;29(1):196-202. doi:10.1377/hlthaff.2009.0077PubMedGoogle ScholarCrossref
4.
Schleifer  D, Silliman  R, Rinehart  C. Still searching: how people use health care price information in the United States. https://www.publicagenda.org/pages/still-searching. Published April 6, 2017. Accessed March 1, 2019.
5.
Hunter  WG, Hesson  A, Davis  JK,  et al.  Patient-physician discussions about costs: definitions and impact on cost conversation incidence estimates.  BMC Health Serv Res. 2016;16(1):108. doi:10.1186/s12913-016-1353-2PubMedGoogle ScholarCrossref
6.
Gordon  DD, Patel  I, Pellegrini  AM, Perlis  RH.  Prevalence and nature of financial considerations documented in narrative clinical records in intensive care units.  JAMA Netw Open. 2018;1(7):e184178. doi:10.1001/jamanetworkopen.2018.4178PubMedGoogle ScholarCrossref
7.
US Census Bureau. S1901: income in the past 12 months (in 2013 inflation-adjusted dollars)—2009-2013 5-Year American Community Survey (ACS). https://data.census.gov/. Accessed March 1, 2019.
8.
Murphy  SN, Mendis  M, Hackett  K,  et al.  Architecture of the open-source clinical research chart from informatics for integrating biology and the bedside.  AMIA Annu Symp Proc. 2007;2007:548-552.PubMedGoogle Scholar
9.
Pedregosa  F, Varoquaux  G, Gramfort  A,  et al.  Scikit-learn: machine learning in python.  J Mach Learn Res. 2011;12:2825-2830. http://www.jmlr.org/papers/volume12/pedregosa11a/pedregosa11a.pdf. Accessed March 1, 2019.Google Scholar
10.
McKinney  W. Data structures for statistical computing in Python. http://conference.scipy.org/proceedings/scipy2010/pdfs/proceedings.pdf. Published 2010. Accessed March 1, 2019.
11.
Jones  E, Oliphant  E, Peterson  P. SciPy: open source scientific tools for Python. https://www.scipy.org/. Published 2019. Accessed April 7, 2019.
12.
Seabold  S, Perktold  J. Statsmodels: econometric and statistical modeling with Python. https://conference.scipy.org/proceedings/scipy2010/pdfs/seabold.pdf. Published 2010. Accessed March 1, 2019.
13.
Shrank  WH, Fox  SA, Kirk  A,  et al.  The effect of pharmacy benefit design on patient-physician communication about costs.  J Gen Intern Med. 2006;21(4):334-339. doi:10.1111/j.1525-1497.2006.00402.xPubMedGoogle ScholarCrossref
14.
Patel  MR, Wheeler  JRC.  Physician-patient communication on cost and affordability in asthma care: who wants to talk about it and who is actually doing it.  Ann Am Thorac Soc. 2014;11(10):1538-1544. doi:10.1513/AnnalsATS.201408-363OCPubMedGoogle ScholarCrossref
15.
Desautels  T, Calvert  J, Hoffman  J,  et al.  Using transfer learning for improved mortality prediction in a data-scarce hospital setting.  Biomed Inform Insights. 2017;9:1178222617712994. doi:10.1177/1178222617712994PubMedGoogle ScholarCrossref
16.
Moriates  C, Soni  K, Lai  A, Ranji  S.  The value in the evidence: teaching residents to “choose wisely.”  JAMA Intern Med. 2013;173(4):308-310. doi:10.1001/jamainternmed.2013.2286PubMedGoogle ScholarCrossref
17.
Osborn  R, Doty  MM, Moulds  D, Sarnak  DO, Shah  A.  Older Americans were sicker and faced more financial barriers to health care than counterparts in other countries.  Health Aff (Millwood). 2017;36(12):2123-2132. doi:10.1377/hlthaff.2017.1048PubMedGoogle ScholarCrossref
18.
Finke  MS, Howe  JS, Huston  SJ.  Old age and the decline in financial literacy  [published online August 24, 2011].  Manag Sci. doi:10.2139/ssrn.1948627.PubMedGoogle Scholar
19.
Mehrotra  A, Dean  KM, Sinaiko  AD, Sood  N.  Americans support price shopping for health care, but few actually seek out price information.  Health Aff (Millwood). 2017;36(8):1392-1400. doi:10.1377/hlthaff.2016.1471PubMedGoogle ScholarCrossref
20.
Chapman  EN, Kaatz  A, Carnes  M.  Physicians and implicit bias: how doctors may unwittingly perpetuate health care disparities.  J Gen Intern Med. 2013;28(11):1504-1510. doi:10.1007/s11606-013-2441-1PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    Health Informatics
    August 30, 2019

    Prevalence of Financial Considerations Documented in Primary Care Encounters as Identified by Natural Language Processing Methods

    Author Affiliations
    • 1Perelman School of Medicine, University of Pennsylvania, Philadelphia
    • 2Mossavar-Rahmani Center for Business and Government, Harvard Kennedy School, Cambridge, Massachusetts
    • 3Center for Quantitative Health, Division of Clinical Research, Massachusetts General Hospital, Harvard Medical School, Boston
    • 4Center for Genomic Medicine, Massachusetts General Hospital, Harvard Medical School, Boston
    JAMA Netw Open. 2019;2(8):e1910399. doi:10.1001/jamanetworkopen.2019.10399
    Key Points español 中文 (chinese)

    Question  How often are cost considerations documented in narrative clinical notes in a primary care setting?

    Findings  In this cohort study of a data set including 222 457 outpatient primary care notes for 46 244 patients at a large academic medical center, 13.1% of patients had at least 1 note indicating a financial conversation with their physician. Specific socioeconomic features were associated with the presence of documented cost considerations.

    Meaning  Although the literature suggests that patients desire to discuss health costs with their physicians, these conversations remain infrequent.

    Abstract

    Importance  Quantifying patient-physician cost conversations is challenging but important as out-of-pocket spending by US patients increases and patients are increasingly interested in discussing costs with their physicians.

    Objective  To characterize the prevalence of financial considerations documented in narrative clinical records of primary care encounters and their association with patient-level features.

    Design, Setting, and Participants  This cohort study applied natural language processing to narrative clinical notes obtained from electronic health records for adult primary care visits. Participants included patients aged 18 years and older with at least 1 primary care visit for an annual preventive examination at outpatient clinics at a US academic health system between January 2, 2008, and July 30, 2013. Data were analyzed in March 2019.

    Main Outcomes and Measures  Presence of financial content documented in narrative clinical notes.

    Results  The data set included 222 457 primary care visits for 46 244 individuals aged 18 years and older; 30 556 patients (60.1%) were female, 27 869 patients (60.3%) were white, and the mean (SD) age was 51.3 (17.7) years. In total, 6058 patients (13.1%) had at least 1 narrative clinical note indicating a financial conversation with their physician. In fully adjusted regression models, the odds of having a financial note were greater among patients with Medicare (odds ratio [OR], 1.27; 95% CI, 1.15-1.41; P < .001) or Medicaid (OR, 1.43; 95% CI, 1.25-1.64; P < .001) insurance, those residing in zip codes with lower median income (OR, 0.97; 95% CI, 0.96-0.98; P < .001), black individuals (OR, 1.40; 95% CI, 1.28-1.53; P < .001), Hispanic individuals (OR, 1.10; 95% CI, 1.01-1.20; P = .03), and those who were unmarried (OR, 1.23; 95% CI, 1.15-1.33; P < .001).

    Conclusions and Relevance  Cost considerations were more likely to be noted in annual preventive examinations than previously observed in intensive care unit admissions, but still infrequently. Associations with particular patient subgroups may indicate differential financial burden or willingness to discuss financial concerns.

    Introduction

    Household out-of-pocket spending on health care has been increasing steadily over recent years in the United States, driven at least in part by the growing popularity of high-deductible health plans.1 In response, physicians are becoming increasingly mindful of patient spending: some physician groups explicitly advocate for the consideration of costs in their clinical guideline documents,2 and a survey3 revealed that 84% of oncologists incorporate patient out-of-pocket costs into treatment recommendations. There is also a growing body of literature examining the prevalence and content of cost-related conversations between patients and physicians. Most US individuals, up to 70% according to a large national study from 2017,4 are interested in discussing cost with their physician, although current estimates of the actual prevalence of cost conversations vary widely, from 4% to 65%.5 These previous estimates largely relied on either survey data, which may be biased, or recorded clinic interactions, which may result in the disruption of care.

    As an alternative to these existing methods, in our previous work,6 we applied machine learning to electronic health records to develop a highly discriminative model identifying the presence and nature of financial considerations in intensive care unit (ICU) clinical notes. Here, we sought to understand the prevalence of such conversations in a cohort likely to be more representative of medicine as a whole, particularly outpatient medicine. We applied the model trained on the ICU data to outpatient primary care notes from a large academic medical center. As in the ICU setting, we also aimed to understand the extent to which these conversations might be associated with patient-level sociodemographic features.

    Methods
    Cohort Generation and Study Design

    We queried the Partners Research Patient Data Registry to identify all adult primary care visits for annual preventive examinations at Massachusetts General Hospital and Brigham and Women’s Hospital between January 2, 2008, and July 30, 2013 (data beyond 2013 were not available because of migration to a new electronic medical record system). To control for the type of visit, we sought to capture routine primary care visits rather than problem-based visits: adult annual preventive examinations were identified using Current Procedural Terminology, Fourth Edition codes 99204, 99205, 99214, and 99215. We identified 222 457 primary care visits for 46 244 individuals aged 18 years and older. In addition to narrative notes, sociodemographic features were extracted from structured data, including age, sex, marital status, race/ethnicity, insurance type, and median household income based on 2013 census data7 imputed from zip code, generating a data mart using i2b2 server software.8

    The Partners HealthCare Human Research Committee approved the study protocol and waived the requirement for informed consent under 45 CFR 46.116 because no participant contact was required in this study based on secondary use of data arising from routine clinical care. This study follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

    Model Training

    We have previously described the derivation of the natural language processing (NLP) model used to classify notes.6 In brief, the notes labeled for the presence of financial conversations from the previous study were used to train a random forest classifier with Python’s scikit-learn package (sklearn.ensemble.RandomForestClassifier, version 0.20.0).9 These notes were randomly split into a training set of 5021 of 5579 notes (90.0%) for model development and tuning, and a testing set of 558 (10.0%) was used to confirm the model’s ability to generalize. The number of notes used to train the model was slightly higher than that in the previous study because nonindex admission notes were also included to allow for more data to train on. As previously discussed,6 standard NLP techniques, including removing punctuation and English-language stop words, stemming, generating unigrams and bigrams, creating a term frequency–inverse document frequency matrix from these unigrams and bigrams, and using principal components analysis to reduce the dimensionality of this matrix, were used to transform the text before model training.

    Areas under the curve calculated by using the trapezoidal rule, precision, and recall were used as the primary evaluation metrics for discrimination on the training and test sets. On the independent testing set, the area under the curve exceeded 0.96. Precision and recall were calculated at different thresholds of the model’s output probability to determine which threshold struck the best balance for our use case and would, therefore, be used on the outpatient data set. In this case, a threshold of 0.70 was chosen for subsequent analysis because it yielded both precision and recall values greater than 0.92 on the testing set.

    Model Evaluation on Outpatient Data Set

    The 222 457 clinical notes in the outpatient data set were preprocessed in exactly the same way as the training set. To validate the performance of the model on the outpatient data set, 1 author manually tagged 100 of these notes to determine a reference standard. On these 100 notes, the area under the curve was 0.83, and precision and recall were 0.83 and 0.93, respectively, at the predetermined 0.70 threshold. One author then manually tagged 300 patient notes identified by the model as having a documented financial conversation into further subcategories according to the nature of the conversation.

    Statistical Analysis

    Summary statistics, including proportion for categorical features and mean (SD) for continuous features, were computed for each group (ie, patients who had engaged in financial conversations and patients who had not) using Python’s pandas (version 0.23.4)10 and scipy (version 1.1.0)11 packages. No data were missing. Comparisons used χ2 tests or single-sample, unpaired, 2-tailed t tests, as appropriate. After these univariate analyses, binomial logistic regression models using Python’s statsmodel (version 0.9.0),12 adjusted for the total number of notes associated with the patient, age, sex, race/ethnicity, insurance type, and zip code median income at earliest recorded visit, were fit with the presence or absence of financial discussion as the dependent variable to estimate the independent effect of each sociodemographic feature in terms of odds ratios (ORs) and 95% confidence intervals. All reported P values are 2-sided, with nominal significance considered to be uncorrected P < .05.

    Results

    The data set as a whole included 222 457 outpatient notes for 46 244 patients seen between January 2, 2008, and July 30, 2013. Of these 46 244 patients, 30 556 (60.1%) were female, 27 869 (60.3%) were white, and the mean (SD) age was 51.3 (17.7) years. In total, 6058 patients (13.1%) were found to have at least 1 note over this period indicating a financial conversation with their physician. For each year between 2008 and 2013, 7.2%, 7.6%, 7.6%, 8.2%, and 6.7% of patients, respectively, had at least 1 cost-related note; each of these percentages is lower than the overall estimate because many patients had multiple years of notes.

    Table 1 describes the sociodemographic features of the study population. In univariate comparisons, compared with those without financial conversations, patients with financial conversations documents in their notes were older (mean [SD] age, 54.1 [16.4] vs 49.9 [17.5] years; t = 19.9; P < .001) and resided in lower-income areas (mean [SD] zip code median income, $70 684 [$28 377] vs $76 782 [$30 366]; t = −16.6; P < .001) than patients without such notes. Financial notes were also more likely to appear in the records of women (4170 of 30 556 patients [13.7%] vs 1888 of 15 688 patients [12.0%]; χ2 = 23.5; P < .001), individuals of Hispanic (1223 of 7776 [15.7%] vs 4835 of 38 468 [12.6%]; χ2 = 56.4; P < .001) and black (975 of 5392 [18.1%] vs 5083 of 40 852 [12.4%]; χ2 = 132.6; P < .001) race/ethnicity, those who were single (1982 of 14 271 [13.9%] vs 4076 of 31 973 [12.8%]; χ2 = 11.2; P < .001), and those with Medicare (643 of 3529 [18.2%] vs 5415 of 42 715 [12.7%]; χ2 = 87.5; P < .001) and non-Medicare (1729 of 8861 [19.5%] vs 4239 of 37 383 [11.6%]; χ2 = 395.2; P < .001) government insurance.

    Multivariable logistic regression was used to determine the independent association of sociodemographic factors with the likelihood of a financial note (Table 2). There continued to be significant associations with neighborhood income (lower zip code median income, OR, 0.97; 95% CI, 0.96-0.98; P < .001), nonwhite race/ethnicity (Hispanic individuals, OR, 1.10; 95% CI, 1.01-1.20; P = .03; black individuals, OR, 1.40; 95% CI, 1.28-1.53; P < .001), marital status (unmarried, OR, 1.23; 95% CI, 1.15-1.33; P < .001), and insurance type (Medicare, OR, 1.27; 95% CI, 1.15-1.41; P < .001; Medicaid, OR, 1.43; 95% CI, 1.25-1.64; P < .001). However, age (OR, 1.00; 95% CI, 1.00-1.00; P = .05) and female sex (OR, 1.05; 95% CI, 0.99-1.12; P = .10) were no longer statistically significantly associated with the presence of financial notes.

    Finally, 300 of the notes identified by the NLP model as having documented financial conversations were then further subcategorized by the nature of the discussion (Table 3); 271 (90.3%) were true discussions of financial matters. Of these notes, 191 (70.5%) included a discussion of health insurance, 132 (48.7%) and 78 (28.8%) documented a change in treatment plan or medication, respectively, and 47 (17.3%) mentioned non–health-related financial concerns (eg, unemployment, legal troubles, and child care expenses).

    Discussion

    In application of a previously validated NLP model to a data set of 222 457 outpatient primary care notes, we found that 13.1% of 46 244 patients had at least 1 documented conversation with their physician that referenced financial considerations. The percentage of patients with at least 1 cost-related conversation in a given year was fairly constant at 6.7% to 8.2% per year. In adjusted analyses, these discussions were more likely for unmarried individuals, individuals residing in areas with lower median household incomes, individuals with government health insurance, and individuals of black and Hispanic race/ethnicity.

    Estimates of the prevalence of financial conversations vary widely across the literature; for a summary, see the article by Hunter et al.5 These estimates include 30% of dialogue transcripts from specialist outpatient visits with a mention of cost,1 15% of primary care physicians in California reporting discussing costs with their patients most or all of the time,13 and 40% of African American women from a single center reporting having a conversation about the cost of their asthma care with their physician.14 In our own previous work6 applying NLP to ICU clinical notes, we reported that 4.2% of patients had at least 1 note reflecting financial considerations during their stay. Clearly, there are many reasons why the reported estimates may differ, including the method of data collection, nature of the clinical context, and criteria used to define a financial conversation.

    Of note, all prior work on this subject is based on either survey data or recorded dialogue from clinic visits. To our knowledge, this study is the first to use NLP to arrive at an estimate of the prevalence of financial conversations in the outpatient setting. That our estimate falls within the range of previously published work not only validates the accuracy of this method but also illustrates how machine learning can be applied to large data sets to answer questions more efficiently and cost-effectively than ever before. Furthermore, we demonstrate how an existing model, in this case an NLP model trained on the ICU notes from our previous work,6 can be repurposed for a second task. This strategy, known as transfer learning, can overcome issues in which a clinical data set is too large to be relabeled entirely (or too small for a model to be trained on it effectively).15 The strong performance of the ICU model on the outpatient visit notes suggests that this model can be successfully applied to other data sets in the future.

    In a random sample of notes with documented financial conversations, we found that 48.7% of conversations resulted in a change to treatment and 28.8% resulted in a change of medication. The finding that patient-physician discussions of cost can result in a change of management is in line with other research suggesting that physicians are increasingly incorporating patient out-of-pocket spending into their decisions-making.2,3,16 Whether these changes are associated with health outcomes remains to be determined.

    Unsurprisingly, our results demonstrate that socioeconomic factors are associated with the presence of cost conversations. Single patients, individuals residing in zip codes with lower median incomes, and those with government insurance including Medicaid likely experience a greater economic burden associated with medical care compared with other groups and, thus, are more likely to discuss cost with their physician. These findings may inform physician behavior and training; for example, medical students and residents may be taught which medications are more or less likely to be approved by public insurance plans. However, the association between Medicare insurance and the presence of financial notes is more surprising, because Medicare is intended to protect senior citizens against medical expenditure risk. We postulate that this may be reflective of global survey findings17 showing that US seniors face more financial barriers to care, despite nearly universal access via Medicare, than their peers in other high-income countries. It may also reflect a decline in financial literacy among seniors, which may increase economic anxiety in this group.18

    We also found that black and Hispanic individuals are more likely to discuss cost with their physicians, as evident in both crude and adjusted models. A recent large patient survey19 found similar results: black and Hispanic individuals were more likely to be aware of price before receiving care, to investigate out-of-pocket spending, and to compare costs across physicians. This result may be confounded by other variables that drive either increased financial burden of care or increased desire to discuss financial concerns for black and Hispanic individuals. On the other hand, we also recognize that if these conversations were primarily initiated by the physician and not the patient, they may reflect unconscious bias by physicians that black and Hispanic individuals have greater economic burden compared with others.20 Drawing a meaningful conclusion from this result is predicated on understanding in greater detail the types of conversations that occur and which party initiated them.

    Our findings add to the existing research exploring the role of financial considerations in encounters between patients and physicians. Continued study of the nature of these conversations can guide future efforts to improve communication between patients and physicians, particularly as a new generation of physicians is being trained.

    Limitations

    We note multiple limitations in this work. Our model was initially trained on inpatient ICU notes, which may involve language and themes that are different from those of outpatient settings. It is also likely that we underestimate the true prevalence of patient-physician financial conversations because some discussions may not be documented in clinical notes or may take place with front desk or billing staff instead of the physician. The present data also do not allow investigation of physician-level features that likely are associated with the probability of such discussions. In addition, other unmeasured variables (eg, medical comorbidities and the presence or absence of an acute problem at time of visit) are likely to be associated with the probability of financial discussion. Also, our analysis was limited to a single academic medical center; future applications of the model to data from other sites may reveal interesting regional and institutional differences in the presence of cost conversations.

    Conclusions

    This study shows the feasibility of transferring a previously developed NLP model to identify documented financial conversations in narrative clinical notes and provides evidence to suggests that these conversations can influence medical decision-making by physicians. The likelihood of documented financial conversations is associated with sociodemographic factors, including marital status, insurance type, and race/ethnicity. This approach adds to the range of strategies for investigating the extent and nature of financial conversations, as a means of better characterizing the financial burden of health care for individual patients.

    Back to top
    Article Information

    Accepted for Publication: July 8, 2019.

    Published: August 30, 2019. doi:10.1001/jamanetworkopen.2019.10399

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Skaljic M et al. JAMA Network Open.

    Corresponding Author: Deborah D. Gordon, MBA, Mossavar-Rahmani Center for Business and Government, Harvard Kennedy School, 79 John F. Kennedy St, Cambridge, MA 02138 (dgordon@mba1999.hbs.edu).

    Author Contributions: Dr Perlis 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.

    Concept and design: Patel, Perlis, Gordon.

    Acquisition, analysis, or interpretation of data: Skaljic, Patel, Pellegrini, Castro.

    Drafting of the manuscript: Skaljic, Patel, Pellegrini.

    Critical revision of the manuscript for important intellectual content: Castro, Perlis, Gordon.

    Statistical analysis: Patel.

    Administrative, technical, or material support: Skaljic, Patel, Pellegrini, Castro, Gordon.

    Supervision: Perlis, Gordon.

    Conflict of Interest Disclosures: Dr Perlis reported receiving personal fees for consulting or for service on scientific advisory boards from Genomind, Psy Therapeutics, RID Ventures, and Takeda and grants from the National Institute of Mental Health and the National Heart, Lung, and Blood Institute outside the submitted work. He also reported holding equity in Outermost Therapeutics and Psy Therapeutics. No other disclosures were reported.

    Disclaimer: Dr Perlis, a JAMA Network Open associate editor, was not involved in the editorial review of or the decision to publish this article.

    References
    1.
    Hunter  WG, Zhang  CZ, Hesson  A,  et al.  What strategies do physicians and patients discuss to reduce out-of-pocket costs? analysis of cost-saving strategies in 1,755 outpatient clinic visits.  Med Decis Making. 2016;36(7):900-910. doi:10.1177/0272989X15626384PubMedGoogle ScholarCrossref
    2.
    Schwartz  JAT, Pearson  SD.  Cost consideration in the clinical guidance documents of physician specialty societies in the United States.  JAMA Intern Med. 2013;173(12):1091-1097. doi:10.1001/jamainternmed.2013.817PubMedGoogle ScholarCrossref
    3.
    Neumann  PJ, Palmer  JA, Nadler  E, Fang  C, Ubel  P.  Cancer therapy costs influence treatment: a national survey of oncologists.  Health Aff (Millwood). 2010;29(1):196-202. doi:10.1377/hlthaff.2009.0077PubMedGoogle ScholarCrossref
    4.
    Schleifer  D, Silliman  R, Rinehart  C. Still searching: how people use health care price information in the United States. https://www.publicagenda.org/pages/still-searching. Published April 6, 2017. Accessed March 1, 2019.
    5.
    Hunter  WG, Hesson  A, Davis  JK,  et al.  Patient-physician discussions about costs: definitions and impact on cost conversation incidence estimates.  BMC Health Serv Res. 2016;16(1):108. doi:10.1186/s12913-016-1353-2PubMedGoogle ScholarCrossref
    6.
    Gordon  DD, Patel  I, Pellegrini  AM, Perlis  RH.  Prevalence and nature of financial considerations documented in narrative clinical records in intensive care units.  JAMA Netw Open. 2018;1(7):e184178. doi:10.1001/jamanetworkopen.2018.4178PubMedGoogle ScholarCrossref
    7.
    US Census Bureau. S1901: income in the past 12 months (in 2013 inflation-adjusted dollars)—2009-2013 5-Year American Community Survey (ACS). https://data.census.gov/. Accessed March 1, 2019.
    8.
    Murphy  SN, Mendis  M, Hackett  K,  et al.  Architecture of the open-source clinical research chart from informatics for integrating biology and the bedside.  AMIA Annu Symp Proc. 2007;2007:548-552.PubMedGoogle Scholar
    9.
    Pedregosa  F, Varoquaux  G, Gramfort  A,  et al.  Scikit-learn: machine learning in python.  J Mach Learn Res. 2011;12:2825-2830. http://www.jmlr.org/papers/volume12/pedregosa11a/pedregosa11a.pdf. Accessed March 1, 2019.Google Scholar
    10.
    McKinney  W. Data structures for statistical computing in Python. http://conference.scipy.org/proceedings/scipy2010/pdfs/proceedings.pdf. Published 2010. Accessed March 1, 2019.
    11.
    Jones  E, Oliphant  E, Peterson  P. SciPy: open source scientific tools for Python. https://www.scipy.org/. Published 2019. Accessed April 7, 2019.
    12.
    Seabold  S, Perktold  J. Statsmodels: econometric and statistical modeling with Python. https://conference.scipy.org/proceedings/scipy2010/pdfs/seabold.pdf. Published 2010. Accessed March 1, 2019.
    13.
    Shrank  WH, Fox  SA, Kirk  A,  et al.  The effect of pharmacy benefit design on patient-physician communication about costs.  J Gen Intern Med. 2006;21(4):334-339. doi:10.1111/j.1525-1497.2006.00402.xPubMedGoogle ScholarCrossref
    14.
    Patel  MR, Wheeler  JRC.  Physician-patient communication on cost and affordability in asthma care: who wants to talk about it and who is actually doing it.  Ann Am Thorac Soc. 2014;11(10):1538-1544. doi:10.1513/AnnalsATS.201408-363OCPubMedGoogle ScholarCrossref
    15.
    Desautels  T, Calvert  J, Hoffman  J,  et al.  Using transfer learning for improved mortality prediction in a data-scarce hospital setting.  Biomed Inform Insights. 2017;9:1178222617712994. doi:10.1177/1178222617712994PubMedGoogle ScholarCrossref
    16.
    Moriates  C, Soni  K, Lai  A, Ranji  S.  The value in the evidence: teaching residents to “choose wisely.”  JAMA Intern Med. 2013;173(4):308-310. doi:10.1001/jamainternmed.2013.2286PubMedGoogle ScholarCrossref
    17.
    Osborn  R, Doty  MM, Moulds  D, Sarnak  DO, Shah  A.  Older Americans were sicker and faced more financial barriers to health care than counterparts in other countries.  Health Aff (Millwood). 2017;36(12):2123-2132. doi:10.1377/hlthaff.2017.1048PubMedGoogle ScholarCrossref
    18.
    Finke  MS, Howe  JS, Huston  SJ.  Old age and the decline in financial literacy  [published online August 24, 2011].  Manag Sci. doi:10.2139/ssrn.1948627.PubMedGoogle Scholar
    19.
    Mehrotra  A, Dean  KM, Sinaiko  AD, Sood  N.  Americans support price shopping for health care, but few actually seek out price information.  Health Aff (Millwood). 2017;36(8):1392-1400. doi:10.1377/hlthaff.2016.1471PubMedGoogle ScholarCrossref
    20.
    Chapman  EN, Kaatz  A, Carnes  M.  Physicians and implicit bias: how doctors may unwittingly perpetuate health care disparities.  J Gen Intern Med. 2013;28(11):1504-1510. doi:10.1007/s11606-013-2441-1PubMedGoogle ScholarCrossref
    ×