Article Text


Patient preferences for venous thromboembolism prophylaxis after injury: a discrete choice experiment
  1. Bryce E Haac1,
  2. Nathan N O'Hara1,
  3. C Daniel Mullins2,
  4. Deborah M Stein1,
  5. Theodore T Manson1,
  6. Herman Johal3,
  7. Renan Castillo4,
  8. Robert V O'Toole1,
  9. Gerald P Slobogean1
  1. 1 School of Medicine, University of Maryland, Baltimore, Maryland, USA
  2. 2 School of Pharmacy, University of Maryland, Baltimore, Maryland, USA
  3. 3 Department of Surgery, McMaster University, Hamilton, Ontario, Canada
  4. 4 Johns Hopkins University Bloomberg School of Public Health, Baltimore, Maryland, USA
  1. Correspondence to Dr. Nathan N O'Hara; nohara{at}


Objective Limited evidence for the optimal venous thromboembolism (VTE) prophylaxis regimen in orthopaedic trauma leads to variability in regimens. We sought to delineate patient preferences towards cost, complication profile, and administration route (oral tablet vs. subcutaneous injection).

Design Discrete choice experiment (DCE).

Setting Level 1 trauma center in Baltimore, USA.

Participants 232 adult trauma patients (mean age 47.9 years) with pelvic or acetabular fractures or operative extremity fractures.

Primary and secondary outcome measures Relative preferences and trade-off estimates for a 1% reduction in complications were estimated using multinomial logit modelling. Interaction terms were added to the model to assess heterogeneity in preferences.

Results Patients preferred oral tablets over subcutaneous injections (marginal utility, 0.16; 95% CI: 0.11 - 0.21, P<0.0001). Preferences changed in favor of subcutaneous injections with an absolute risk reduction of 6.98% in bleeding, 4.53% in wound complications requiring reoperation, 1.27% in VTE, and 0.07% in death from pulmonary embolism (PE). Patient characteristics (sex, race, type of injury, time since injury) affected patient preferences (P<0.01).

Conclusions Patients preferred oral prophylaxis and were most concerned about risk of death from PE. Furthermore, the findings estimated the trade-offs acceptable to patients and heterogeneity in preferences for VTE prophylaxis.

  • thromboembolism
  • anticoagulation
  • trauma management
  • adult surgery

This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See:

Statistics from

Strengths and limitations of this study

  • This study quantifies patient preferences for venous thromboembolism prophylaxis in a high-risk and often difficult to research population.

  • The results provide valuable benefit–risk trade-offs estimates to guide clinicians in a common decisional dilemma.

  • High face validity in the choice sets is demonstrated by the directionality, magnitude and consistency of the responses.

  • The high response rate captured in this prospective study reduces response bias present in other survey methods.

  • The choice sets presented to respondents were hypothetical scenarios, and the respondent’s actual choices may be different.


Traumatic injury is a well-described risk factor for the development of venous thromboembolism (VTE). The incidence of VTE among trauma patients ranges from 20% to 90% without any preventative measure.1 In addition, pulmonary embolism (PE) is the third most common cause of death in patients who survive the first 24 hours following injury.1–4 Orthopaedic trauma patients in particular have several well-known risk factors for VTE placing them at exceptionally high risk.2 5–8 Fortunately, chemoprophylaxis has been shown to significantly reduce the incidence of VTE in this population.9 However, controversy exists as to the optimal VTE prophylaxis regimen in orthopaedic trauma patients.10–15

For many orthopaedic populations, the American College of Chest Physicians (CHEST) and the Eastern Association for the Surgery of Trauma recommend enoxaparin (low-molecular-weight heparin (LMWH)) by subcutaneous injection for VTE prophylaxis, but recent studies show that acetylsalicylic acid (aspirin), an oral tablet, may be an equally effective alternative with lower risk of bleeding complications.10–15 However, only limited data are available specific to orthopaedic trauma patients who may have even higher risk for both VTE events and bleeding.16 The Orthopaedic Trauma Association Evidence Based Quality Value and Safety Committee highlights variability in prescribed regimens due to the poor scientific support for various regimens and emphasises the need for guidelines to improve patient care.17

The CHEST guidelines emphasise the need for systematic reviews of patient values and preferences when creating guidelines for specific populations.18 Creation of guidelines requires making risk and benefit trade-offs, and patient values regarding VTE prophylaxis depend on the health outcomes considered. Furthermore, defining the heterogeneity of preferences in this patient population is necessary to provide valuable individualised VTE prevention options. Implementing guidelines that consider patient preferences may increase patient satisfaction with and improve adherence to clinical treatments.19 Patient medication refusal is a leading cause of non-administration of VTE prophylaxis in inpatients, and missed doses are highly associated with increased VTE incidence.20–22 In a study of medical and surgical patient preferences for VTE prophylaxis regimens, the majority of patients preferred oral administration over subcutaneous injection if all other factors were equal.23 Patients who preferred subcutaneous administration presumed a faster onset of action and were less likely to refuse administration.

Existing VTE prevention studies do not evaluate patient preferences, investigate acceptable trade-offs of the risks and benefits of those medications or determine heterogeneity in preferences based on demographic and clinical characteristics. The purpose of this study was to elicit the preferences of orthopaedic trauma patients toward currently available VTE prophylaxis, examine acceptable trade-offs of the potential complications related to those mediations and determine heterogeneity in preferences among patient subgroups.


A discrete choice experiment (DCE) was prospectively administered to orthopaedic trauma patients at a level 1 trauma centre. DCEs are a quantitative technique used to measure individual preferences in a variety of healthcare settings by administering surveys that ask individuals to choose the best option between two or more hypothetical scenarios or choice sets.24–26 Options are described with a fixed set of attributes levels that vary in each scenario. The data collected can be used to assess the relative importance of each attribute and acceptable trade-offs among attributes. An estimate of preference can be described as the marginal utility for a given attribute level. Marginal utility can be positive or negative, with numbers farther from zero indicating a stronger preference. Monetary costs can be included to produce willingness-to-pay estimates.

Study setting and population

This study was conducted at the R Adams Cowley Shock Trauma Centre in Baltimore, Maryland, and received prior approval by the institutional review board at the University of Maryland School of Medicine. All adult (> 18 years) patients treated with pelvic or acetabular fractures or an operative extremity fracture were assessed for eligibility from November 2015 to February 2016. Patients who were unable to consent due to intubation or altered mental status and non-English-speaking patients were excluded. On written consent, patients were enrolled in the study as inpatients or at an outpatient follow-up appointment within 4 months from their initial admission for their injury.

Study design

The attributes and their corresponding levels were selected based on a literature review, patient interviews, expert consultation and a retrospective review of patient outcomes. Medication attributes used in the DCE included medication administration route (oral tablet vs subcutaneous injection), cost, possible side effects including bruising or stomach pain, risk of having a bleeding complication that requires a blood transfusion, risk of having a wound complication that requires another operation, risk of VTE requiring therapeutic anticoagulation for 6 months and risk of death due to PE. These attributes were chosen to reflect medication qualities that patients are aware of when taking medications (route, cost, side effects) and clinically important outcomes. Values for these attributes were based on available literature and clinical experience with two commonly prescribed VTE prophylaxis medications in this population: LMWH (a subcutaneous injection) and aspirin (an oral tablet). Attributes were not reflective of other oral anticoagulants because those medications are typically used for treatment of VTE events rather than prevention and the focus of this DCE is preferences for prophylaxis administered to prevent VTE events.

Forty choice sets were developed using a Bayesian D-optimal design with JMP V.12 software (SAS Institute, Cary, North Carolina, USA) to ensure maximum variation in attribute comparison. The 40 choice sets were then randomly divided into four surveys, each with 10 choice sets, to minimise respondent burden. As documented by Sandor et al,27 using heterogeneous designs produce substantial improvements in efficiency over a single survey and provide more precision in estimating true parameters. Each choice set compared two hypothetical VTE prophylaxis medications described by their attributes (figure 1). Patients were randomly assigned one of the four self-administered surveys. A member of the research staff was available for questions as the study participant completed the survey. Demographic data including age, sex, race (as defined by the participant), type of injury, Injury Severity Score, American Society of Anesthesiologists (ASA) physical status, income, health insurance status, days on prophylaxis and timing of recruitment (inpatient vs outpatient) were collected from both the survey and the medical record. The type of VTE prophylaxis was not collected as part of the study. However, at the time of the study, VTE prophylaxis by LMWH was the standard hospital protocol, and it is reasonable to assume that this was prescribed to all study participants unless there was a contraindication.

Figure 1

Sample question from the discrete choice experiment survey administered to participants. In each question, the values for each hypothetical medication are varied.

The target sample size for this study was derived by the Rule of Thumb calculation described by Orme and our a priori decision to conduct multiple subgroup analyses.28 Based on this calculation,28 we determined that a sample of 25 study participants would be required in each possible subgroup category for adequate statistical power. Given known proportions of admission data for this population, a sample size exceeding 200 participants was required to adequately assess heterogeneity in preferences, particularly on sex, race and health insurance status.

Data analysis

Data collected through the DCE survey allow the quantification of and statistical inference about the relative importance of VTE prophylaxis medication attributes. A multinomial logit model,29 30 with effect coding, was used to estimate patient preferences using marginal utility, willingness to pay (WTP) and acceptable trade-off estimates for a 1% reduction in VTE complications or side effects. Marginal utility is a measure of patient preference, with the estimate signifying the strength and direction of one’s preference toward the attribute. With this analysis, we are able to determine the relative magnitude of patient preferences to avoid VTE-related complications in association with their medication choice. Preference heterogeneity was subsequently assessed by adding an interaction term into the model with a priori determined variables of interest. These variables included age (categorised as <40, 40–59, >60), sex, race, ASA status (≤2 vs >2), the location of primary injury (upper extremity vs lower extremity), household income (categorised as ≤$20 000, $20 000–$49 999, $50 000–$74 999, ≥$75 0000), health insurance status (any vs none) and the location of recruitment. All data analysis was conducted using the Choice Modelling platform in JMP V.12.


Of the 310 patients screened for participation, 50 were ineligible (40 unable to consent due to altered mental status, eight non-English speaking, two contraindicated for VTE prophylaxis) and 28 (11%) patients refused participation. Of the 232 patients included in the analysis, the mean age was 47.9 years, with 56.9% male, and 66.8% were white. The majority of participants had a lower extremity injury (83.6%), with a mean injury severity score of 11.7, and were fully insured (83.1%) (table 1).

Table 1

Characteristics of orthopaedic fracture participants (n=232)

Patients most strongly preferred a reduction in risk of death by PE (marginal utility, 4.57; p<0.0001), distantly followed by a reduction in the risk of VTE requiring therapeutic anticoagulation, wound complications requiring another surgery and bleeding complications requiring a transfusion (table 2). Patients were willing to pay $1686.90 for a 1% absolute reduction in risk of death due to PE compared with $92.29 or less for a 1% absolute reduction in any of the other measured outcome variables. Patients also preferred to take oral tablets (marginal utility, 0.16; p<0.0001) and were willing to pay $117.45 to receive prophylaxis via oral route over subcutaneous injection. Possible medication side effects, such as stomach pain or bruising, did not significantly influence patient preferences (p>0.1).

Table 2

Patient preferences and valuation of venous thromboembolism prophylaxis attributes

To change patient preference in favour of subcutaneous injections requires a 6.98% absolute reduction in the risk of bleeding complications requiring transfusion, a 4.53% absolute reduction in the risk of wound complications requiring reoperation and a 1.27% absolute reduction in risk of VTE requiring therapeutic anticoagulation (table 3). In contrast, only a 0.07% absolute reduction in risk of death due to PE was needed to change patient preference.

Table 3

The absolute risk reduction (ARR) of a potential complication that a patient would be willing to accept to change their route preference from oral to subcutaneous injection prophylaxis

In our subgroup analyses examining heterogeneity in preferences, patients who were female, white or had lower extremity injuries demonstrated significantly stronger preference for oral VTE prophylaxis over subcutaneous injections (p<0.05) (table 4). Patients with upper extremity injuries valued a reduction in risk of bleeding complications more than patients with lower extremity injuries (p=0.01). Patients who were recruited as an inpatient valued a reduction in risk of wound complications requiring reoperation more than patients who were recruited from the outpatient clinic (p<0.01). There were no other significant associations between the tested covariates and our included VTE prophylaxis attributes.

Table 4

Subgroup analysis quantifying heterogeneity in patient preferences


Consistent with previous studies,23 our study demonstrates a strong patient preference for oral VTE prophylaxis over subcutaneous injection when all other relevant attributes are equal. However, patients only required a small reduction in the absolute risk of death due to PE to change their preference in favour of a subcutaneous injection. When choosing between VTE prophylaxis regimens, patients most valued (in order): risk of death due to PE, risk of VTE requiring therapeutic anticoagulation, risk of wound complications requiring reoperation and risk of bleeding complications requiring transfusion. A defined reduction in any of these outcomes could change patient preference to favour the subcutaneous injection route. In addition, underlying patient factors such as sex, race, type of injury and inpatient status led to significant heterogeneity in patient preferences.

To our knowledge, our study is the first to assess the weight of patient-valued outcomes regarding potential risks and complications of VTE prophylaxis. The study also determined patient characteristics associated with heterogeneity in their preferences. Previous studies have shown that patient refusal is a common reason for missed VTE doses and increases the risk of a VTE event.20–23 Patient preference for oral medications is also well documented.23 31 Our study also demonstrates a strong preference for oral medications, but our results show that this preference can change if the risk of the aforementioned patient-important outcomes is high enough. Patients were particularly concerned about the risk of death due to PE, requiring only a 0.07% absolute reduction in risk of death to change patient preference in favour of a subcutaneous injection. In addition, only relatively small reductions in the risk of other outcomes were required to change patient preference. Furthermore, the preference for route varied significantly depending on the patient’s sex, race, type of injury and inpatient status.

While the design of the DCE enables the assessment of risk–benefit trade-offs among subgroups, it does not allow for qualitative analysis of patient preferences. As a result, we are only able to speculate as to why patients valued certain outcomes more than others. In addition, the choice sets were hypothetical scenarios, and patient’s actual choices may be different. The greater value placed on risk of death due to PE compared with other outcome measures could be a result of death being the easiest outcome variable for the average patient to understand. We were unable to control for patient disposition in our analysis (home vs rehab), but we did compare responses of inpatients to outpatients. Patients who were recruited as an inpatient were more concerned about the risk of reoperation than patients recruited as outpatients, potentially because their injury and initial operation were more recent in their memory. In addition, study participants had varying lengths of VTE prophylaxis prescribed at time of recruitment, and some patients were closer to time of injury and initial operation than others, although, when assessed, time since injury did not affect patient preferences.

Some participants had personal experience with one or more of the measured outcomes, while others had no history of complications, which we were unable to control for in our final analysis. In the same manner, the mean Injury Severity Score (ISS) of our sample was 11.7, likely as a result of many patients having isolated orthopaedic injuries as well as more severely injured patients not having the mental capacity to complete the survey. ISS ranged from 4 to 34, but there is the possibility that our results may suffer from some respondent bias if trying to extrapolate to a more severely injured population. Lastly, we did not collect data on patient education level which could affect the patient’s understanding of certain outcomes; however, income and insurance level may be surrogate markers for education and were included in the analysis.

In the current era of patient-centred healthcare, it is important that we consider all outcomes that patients value and the heterogeneity in those preferences when conducting clinical comparative effectiveness research and when making clinical guidelines in order to improve healthcare delivery and reduce cost.19 32 33 Our data demonstrate that orthopaedic trauma patients prefer VTE prophylaxis by oral tablet to prophylaxis by subcutaneous injection when all other relevant attributes are equal. However, the risk of death due to PE is the dominant concern when choosing a regimen. Our study is the first to document the value patients place on various clinically important outcomes related to VTE prophylaxis. In addition, we define the underlying patient factors that contribute to variation in VTE prophylaxis preferences with risk–benefit trade-offs among subgroups in this important area of ongoing debate. In the era of patient-centred healthcare, future studies and clinical guideline recommendations comparing available VTE prophylaxis regimens should focus on the outcomes most important to patients and incorporate patient trade-off estimates to ensure their work is reflective of patient preferences.


The authors thank all of the respondents for their time and participation in the study.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
View Abstract


  • Contributors BH contributed to the literature search, study design, data collection, data interpretation, writing and critical revision. NNO contributed to the literature search, study design, data analysis, data interpretation, writing and critical revision. CDM contributed to the data interpretation and critical revision. DS, TTM, HJ, RVO and GPS contributed to the literature search, study design, data interpretation and critical revision. RC contributed to the study design, data analysis, data interpretation and critical revision. All authors have approved the final version of the article submitted and agree to be accountable for all aspects of the work.

  • Funding Research reported in this manuscript was funded through a Patient-Centered Outcomes Research Institute (PCORI) Award (PCS-1511-32745). The views in this publication are solely the responsibility of the authors and do not necessarily represent the views of the PCORI, its board of governors or methodology committee.

  • Competing interests CDM reports consulting with Bayer, Daiichi Sankyo, Janssen/J&J, Mundipharma, NovoNordisk and Pfizer and receiving grants from Bayer, Novartis, Merck and Pfizer. TTM receiving grants from the US Air Force and serves as an advisor for Decisio Health. TTM reports consulting with Stryker, Globus and Smith & Nephew, being paid for expert testimony from various law firms and payment for lectures by the Maine Review Course. RVOT reports consulting with Coorstek (Zimmer) and Smith & Nephew and receiving royalties from Coorstek. GPS reports payments for presenting by Zimmer Biomet. No other disclosures were reported.

  • Patient consent Obtained

  • Ethics approval University of Maryland Institutional Review Board.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement No additional data are available.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.