From the *Duke University School of Medicine; the †Fuqua Schoolversity; the ‡Division of Thoracic and Cardiovascular Surgery; and tMaxillofacial, and Oral Surgery, Duke University Medical Center, Dur

Received for publication April 27, 2017; accepted in revised form D

No benefits in any form have been received or will be received relatto the subject of this article.

678.e1 r � 2018 ASSH r Published by Elsevier, I

SCIENTIFIC ARTICLE

Conjoint Analysis of Treatment Preferences for

Nondisplaced Scaphoid Fractures

Ronnie L. Shammas, BS,* Nathan Mela,† Scott Wallace, BS,† Betty C. Tong, MD, MHS,‡Joel Huber, PhD,† Suhail K. Mithani, MD§

Purpose We used conjoint analysis to assess the relative importance of factors that influence apatient’s decision between surgical or nonsurgical management of a nondisplaced scaphoidfracture. Our hypothesis was that out-of-pocket costs will have a greater influence on decisionmaking than the time spent in a cast or brace, degree of soreness, or the risk of treatment failure.

Methods Two-hundred and fifty participants were recruited using Amazon Mechanical Turk andasked to assume that they had experienced a nondisplaced scaphoid waist fracture. They thenindicated their relative preferences among 13 pairs of alternatives with variations in the followingattributes: time in a cast, time in a brace, duration of ongoing soreness, risk of treatment failure (bywhich we meant scaphoid nonunion), out-of-pocket costs based on estimates of direct costs($500e2,500), and apprehension about surgery. A conjoint analysis was used to determine therelative importance of these factors when choosing between surgical or nonsurgical management.

Results The factor with the greatest influence on treatment choice was the cost of the pro-cedure. After assessing the respondent’s apprehension to undergo surgery, a sensitivityanalysis showed the proportion of respondents who would choose surgery given differentoutcomes. To make the predicted share of those who are “not worried” about surgery equal tothose who are “somewhat worried” or “a little worried” would require that the cost of surgeryincrease by $2,700. In addition, 2 weeks in a cast, 3 weeks in a brace, 2 months of soreness, ora 2% increase in the risk of fracture nonunion generates the same surgical choice probabilityas a $2,000 increase in the out-of-pocket cost of surgery.

Conclusions As conceptualized in this conjoint analysis, out-of-pocket costs and apprehensionabout surgery seem to have a greater impact on a decision for surgery than the time spent in abrace or cast and the risk of treatment failure. (J Hand Surg Am. 2018;43(7):678.e1-e9.Copyright � 2018 by the American Society for Surgery of the Hand. All rights reserved.)

Type of study/level of evidence Economic and decision analysis III.Key words Conjoint analysis, hand surgery, patient preferences, scaphoid fracture, shareddecision making.

of Business, Duke Uni-he §Division of Plastic,ham, NC.

ecember 20, 2017.

ed directly or indirectly

Corresponding author: Suhail K. Mithani, MD, Division of Plastic and ReconstructiveSurgery, Department of Surgery, Duke University Medical Center, Box 3945, Durham,NC 27710; e-mail: suhail.mithani@duke.edu.

0363-5023/18/4307-0016$36.00/0https://doi.org/10.1016/j.jhsa.2017.12.021

nc. All rights reserved.

FOR SCAPHOID FRACTURES 678.e2

P ATIENTS WHO EXPERIENCE nondisplaced scaphoidfractures may be treated with cast immobili-zation or surgical treatment with various types

of internal fixation. Immobilization of the wrist for atleast 6 to 12 weeks is an effective treatment withbony union achieved in greater than 90% ofpatients.1e3 However, cast immobilization iscumbersome and may lead to temporary stiffening ofthe wrist, reduced grip strength, a longer return timeto manual work, and a prolonged healing time.3e5

Internal fixation limits immobilization and providespatients with an earlier return to work and higherrates of union.5 However, patients undergoing sur-gery may be at an increased risk for osteoarthritis,soft-tissue injury, and implant-related complications.5

Furthermore, the risks associated with surgery willvary according to the type of internal fixationselected. Thus, the best treatment for nondisplacedscaphoid fractures remains disputed.

The importance of patient preferences in shareddecision making has gained increased attention, andbecomes particularly valuable when many effectivetreatment options exist.6e14 Often, a person’s valuesmay influence his or her treatment preferences. Phy-sicians can weigh patient preferences in the context ofthe available clinical evidence to facilitate shareddecision making. Conjoint analysis is one methodthat has been successfully used to develop outcomemeasures and to study how a person’s values mayaffect his or her treatment preferences.6,15e18

Conjoint analysis is based on the premise that eachtreatment derives value from its expected advantagesand disadvantages. It provides estimates of what arethe most important factors to patients when decidingamongst treatments.7,15e22 Furthermore, when out-of-pocket cost is incorporated as a study attribute,the conjoint analysis can provide an estimate abouthow much a person is willing to pay for a change in agiven attribute.23e26 In this study, we tested the hy-pothesis that out-of-pocket costs would have a greaterinfluence on decision making for the management ofa scaphoid fracture when compared with the timespent in a cast or brace, degree of soreness, or the riskof treatment failure. We also determined the relativeimportance of each attribute as it relates to a surgicalor nonsurgical treatment.

PATIENT PREFERENCES

MATERIALS AND METHODSThe Institutional Review Board granted an exemptresearch status to this protocol. The authors of thisstudy selected 5 attributes that were deemed to in-fluence a person’s quality of life after a wrist fracture.

J Hand Surg Am. r V

These included time in a cast, time in a brace,remaining soreness and stiffness, risk of treatmentfailure, and out-of-pocket cost. We also assessed anindividual’s level of apprehension about surgery. Theattributes selected for this study are those that havebeen consistently addressed in research and patientcare. Our survey was developed based on the inter-pretation of the best available evidence from ran-domized clinical trials that compare castimmobilization and screw fixation for a nondisplacedscaphoid fracture, with the understanding that thereare various interpretations of the best available evi-dence (Table E1, available on the Journal’s Web siteat www.jhandsurg.org).5,27e33 This method of attri-bute selection has been employed in previousstudies.34e36 The various levels for each attributereflect current practice patterns and data as describedin the literature. By providing different levels of eachattribute for 13 hypothetical relative preference ex-periments, it is possible to estimate a value of eachlevel for each respondent. The levels assigned to eachattribute are presented in Table E1, and are as fol-lows: time in a cast (2, 4, or 8 wk), time in a brace (2,4, or 8 wk), remaining soreness or stiffness (2, 4, or 6months), risk of treatment failure (3%, 5%, or 10%),and out-of-pocket costs ($500, $1,000, or $2,500). Acomplete example of the administered survey may befound in Appendix A (available on the Journal’sWeb site at www.jhandsurg.org).

Participants and survey process

Participants of this survey were recruited from anonline panel of members using Amazon MechanicalTurk (MTurk) to administer an electronic survey.37

This provides a large pool of users who can berecruited for academic or private sector research sur-veys. The MTurk method of data collection has beenpreviously validated in obtaining high-quality data inan inexpensive and rapid manner.6,7,10,12,19,22,34,38e43

Similar papers have suggested that a sample size canbe based on prior studies.42 On the basis of the limitednumber of reports in the medical6,7,21,34,41 andorthopedic literature,19,42 the authors concluded thatthe appropriate sample size for this study wasapproximately 250 surveys. Furthermore, beforereleasing the finalized survey, a separate pilot test ofthe survey was conducted to provide assurance that thesurvey was well comprehended and that a sample of250 responses would produce sufficiently small stan-dard errors and stable measures of the importance ofthe tested surgical procedures. Using the onlineMTurk survey, introductory questions first assessedsimilar past injuries, relative activity levels, types of

ol. 43, July 2018

FIGURE 1: Representative example of 1 of 13 hypothetical choice experiments that the survey respondents were asked to answer.

FIGURE 2: Relative importance of the 5 attributes, scaled to sumto 100%. The values listed are representative of the averagevalues across respondents. The majority of respondents deemedcost to be the most important attribute, followed by the risk oftreatment failure, length of soreness, time in a cast, and then timein a brace.

678.e3 PATIENT PREFERENCES FOR SCAPHOID FRACTURES

work (eg, manual), and handedness. Surveyrespondents were then asked how they feel about thepossible outcomes (time in a cast, time in a brace,remaining soreness and stiffness, risk of treatmentfailure, and cost) for the treatment of this injury to theirdominant hand.

Participants were then asked to imagine havingrecently experienced a wrist fracture, and to considertheir reaction to a series of possible outcomes andtreatment options. The preference experiment pre-sented each participant with 13 hypothetical com-parisons that were generated by combining variouslevels of the 5 attributes of interest in an AdaptiveConjoint Analysis using Sawtooth Software.44

Figure 1 gives an example of one of the preferencechoice tasks. Each pair of choice options wasdesigned algorithmically for each participant tomaximize the information gained about each partici-pant’s preferences through a limited number of re-sponses. Based on how each person responded to the13 hypothetical comparisons, it is possible to estimatethe relative importance of the individual attributesthat best reflects the judgments for each respondent.44

After making his or her relative preferences, eachrespondent was then shown a graph (Fig. 2) that re-flected the importance of each attribute in his or herindividual judgments. Respondents indicated theirreaction to this graph to assess how well the conjointsurvey reflected their individual values.

The final portion of the survey reflected a con-versation that a patient may have with his or hersurgeon when discussing the treatment options for hisor her injury (Appendix A, available on the Journal’sWeb site at www.jhandsurg.org). After beinginformed and educated about both surgical andnonsurgical options, the respondents indicated theirlevel of apprehension for undergoing surgery (1 ¼not worried, 2 ¼ a little bit worried, 3 ¼ somewhatworried, and 4 ¼ very worried). This question isimportant as it provides an assessment of apprehen-sion about surgery that is independent of the costs

J Hand Surg Am. r V

and benefits that were assessed earlier. Finally, therespondents were asked to make their decision be-tween surgery and no surgery as is shown in Figure 3.

Statistical analysis

Results of the conjoint analysis were calculated withSawtooth Software, using hierarchical Bayesianmodeling to generate the individual-level conjointutilities that best reproduced each respondent’schoices.15,45 These conjoint utilities provided a basisfor distinguishing respondents’ feelings about indi-vidual aspects of potential treatment options, for

ol. 43, July 2018

FIGURE 3: Final decision-making task. Respondents were asked to make a final decision between undergoing surgery and no surgeryfor the treatment of their nondisplaced scaphoid fracture.

PATIENT PREFERENCES FOR SCAPHOID FRACTURES 678.e4

identifying the relative importance of each testedattribute, and for projecting likely reactions to otherplausible treatment options (eg, an increase in theout-of-pocket cost of surgery). Conjoint parameterestimates were exported and analyzed alongside othersurvey data. The preference values obtained from the13 pairwise judgments did not directly account for anemotional response a patient might have had for thespecific surgery shown in Figure 3. Thus, we built alogistic model that predicted an individual’s choice toundergo the surgery as a function of the individualutility difference for the 5 attributes, and included 3variables to reflect the 4 levels of an individual’s levelof apprehension for undergoing surgery. This modelallowed us to estimate the percent of respondentswho would choose surgery for cases that differedfrom those provided in Figure 3, and for respondentswho had varying levels of apprehension. The logisticregression model could then be used to generate ananalysis that projects the percent of respondents whowould choose to undergo surgery under variousattribute conditions. In particular, the sensitivityanalysis is able to estimate the change needed foreach attribute to compensate for a $500 increase inout-of-pocket costs.

RESULTSA total of 250 individuals participated in the survey.The demographic and individual characteristics ofthe cohort are shown in Table 1. Although re-spondents were not drawn directly from a patient

J Hand Surg Am. r V

population, they reported moderate familiarity withinjuries of this type. In the sample, 21% (n ¼ 53)said that they had previously broken their wrist, 40%(n ¼ 100) said that they have had a different wristinjury that interfered with their normal activities,14% (n ¼ 36) had broken a different part of their armor hand before, and 16% (n ¼ 39) had broken a boneoutside of their arm/hand. However, 39% (n ¼ 90)reported that they had never experienced an injurylike that described in the scenario. In addition, 47%(n ¼ 118) of respondents denoted that they hadpreviously undergone some type of surgery. Twenty-five percent (n ¼ 63) of respondents reported thattheir work involved regular manual labor or heavylifting, and the sample was generally reflective ofnational norms for income, education, and employ-ment status.

The graph in Figure 2 reflects how important eachattribute was across respondents. The relativeimportance of an attribute to an individual respondentvaried, with a standard deviation of approximately 10percentage points across the sample cohort. Subjectsrated cost of treatment as the most important attributein decision making, followed by the risk of treatmentfailure, length of remaining soreness or stiffness, timein a cast, and time in a brace. Overall, 48% (n ¼ 121)of respondents indicated that the conjoint survey re-flected their views “Very well,” 42% (n ¼ 105) said“Well,” 9% (n ¼ 23) said “OK,” and 0.4% (n ¼ 1)said “Poorly.” As shown in Figure 4, the averageutilities of the 5 attributes had unequal weight, andwere relatively linear when compared within the

ol. 43, July 2018

TABLE 1. Demographic Information

Characteristic n ¼ 250 Percentage

Age, y

18e25 35 14

26e30 83 33.2

31e35 55 22

36e40 44 17.6

41e50 18 7.2

51e60 12 4.8

61e70 3 1.2

Sex

Male 143 57.2

Female 107 42.4

Race

White 174 69.6

African American 28 11.2

Other 48 19.2

Handedness

Right 221 88.4

Left 29 11.6

Education level

High school/GED 31 12.4

Some college 102 40.8

Four-year college degree 97 38.8

Master’s degree 16 6.4

Doctoral degree 4 1.6

Annual household income

Less than $20,000 34 13.6

$20,000e39,999 89 35.6

$40,000e59,999 57 22.8

$60,000e79,000 41 16.4

$80,000e99,999 15 6.0

$100,000 or more 14 5.6

Employment status

Full-time employed 169 67.6

Part-time employed 39 15.6

Unemployed seeking work 16 6.4

Disabled 6 2.4

Retired 1 0.4

Other 19 7.6

678.e5 PATIENT PREFERENCES FOR SCAPHOID FRACTURES

levels of each attribute. Furthermore, the individualpreferences did not systematically differ based on arespondent’s familiarity with the injury.

With regard to apprehension about surgery, of the250 respondents, 27 said that they would be “notworried” about surgery. Of those, 70% (n ¼ 19) chosethe surgical option. Of the 89 respondents who said that

J Hand Surg Am. r V

they were “a little bit worried,” 42% (n ¼ 37) chosesurgery. Of the 90 respondents who said that they were“somewhat worried,” 28% (n ¼ 25) chose surgery.Lastly, of the 44 respondents who said that they were“very worried,” 27% (n ¼ 12) chose surgery. To ac-count for differences in apprehension about surgery, webuilt a logistic regressionmodel that predicted choice ofsurgery from the options shown in Figure 3 as a func-tion of the individual conjoint utilities plus coefficientsfor the 4 levels of apprehension about surgery. Thisregression model then provided an estimate regardingthe effect of out-of-pocket cost on an individual’sapprehension about surgery. We then determined whatchanges in out-of-pocket cost would be necessary topredict a choice of surgery for 50% of each apprehen-sion level. Thus, for those who are “not worried,” anadditional cost of $1,300 would reduce their aggregatechoice share from 70% to 50%. For those who are “alittle bit worried,” out-of-pocket costs would have to bereduced by $500. Finally, for those are “somewhatworried” or “very worried,” a reduction of $1,400 inout-of-pocket costs is needed for 50% to choose sur-gery. For the other attributes, it is possible to ask whatchanges are needed to compensate for the choice shareeffects of an additional $500 payment. To avoid apayment of an additional $500, the model predicts thatthe average respondent will accept 2 weeks in a cast, 3weeks in a brace, 2months of soreness, or a 2% increasein the risk of treatment failure.

DISCUSSIONThis study offers an initial perspective on individualpreferences when deciding between treatment optionsfor a nondisplaced scaphoid fracture. The results ofthis study suggest that people deem the cost of theprocedure and their apprehension about surgery asthe most important factors in their decision making.

This survey did not assess the societal costs thatare associated with each treatment, but ratherassessed the impact of out-of-pocket costs. Previousstudies have shown that out-of-pocket costs are apowerful determinant in patient decision making.19,46

O’Hara et al19 found that rather than pay an addi-tional $1,000 for a total shoulder arthroplasty, re-spondents preferred to drive more than 7 hours orwait more than 13 months for surgery. Interestingly,in a study by Kim et al42 that examined the prefer-ences of patients scheduled for carpal tunnel releaseusing conjoint analysis, the authors found that med-ical costs were the least important attribute in patientdecision making. As the authors suggest, this findingis most likely due to the benefits afforded by the

ol. 43, July 2018

FIGURE 4: The average utilities of the 5 attributes. As depicted in this graph, the average utilities of the 5 attributes have unequalweight, and are relatively linear when compared within the levels of each attribute.

PATIENT PREFERENCES FOR SCAPHOID FRACTURES 678.e6

National Health Insurance system in Korea.42 This isimportant to acknowledge when considering theapplicability of this study’s results to countries otherthan the United States where health care costs arecovered by an individual’s employer. Because mostpatients are covered by health insurance, out-of-pocket costs have a reduced influence as comparedwith a privatized system such as the United States.Therefore, this study’s results regarding the influenceof out-of-pocket costs may not be generalizable toother countries.

Apprehension about surgery was the second mostimportant factor in a person’s decision making.Although this topic has yet to be explored withinhand surgery, multiple studies have stressed thatpreoperative education can help reduce patientapprehension.47e49 This is important because appre-hension about surgery not only influences patientdecision making, but also strongly interacts withpostoperative outcomes.50 By educating patientsabout the risks and benefits of their treatment, sur-geons may be able to alleviate the apprehensionassociated with the concept of surgery to helpimprove postoperative outcomes. However, there aremany reasons a patient may be apprehensive aboutsurgery. Thus, we cannot determine from these re-sults the exact degree to which apprehension aboutsurgery and its individual causes correlate with aperson’s decision making.

There are several limitations to this study. First,respondents were recruited using Amazon MTurk,potentially introducing selection bias. Prior studies inthe surgical literature have used Amazon MTurk as anindirect assessment of patient preferences in medical

J Hand Surg Am. r V

decision making. Tong et al6 and more recentlyStreufert et al44 used Amazon MTurk in conjunctionwith conjoint analysis to assess patient preferences inlung cancer treatment and first-time anterior shoulderdislocation, respectively. Nonetheless, injury andtreatment of our respondents is hypothetical and is aninherent form of bias in this study. We cannot say withcertainty that a patient who has sustained this injurywill respond in the same manner as a respondentrecruited from Amazon MTurk. Furthermore, a limi-tation of this study design is that the attributes selectedfor inclusion were not prescreened in a representativepatient focus group. Instead, the authors selected at-tributes and levels based on available clinical evi-dence, as has been previously done.34e36 In addition,the practice patterns presented to respondents partiallyreflect those of the hand surgeons at our institution, andmay differ from those of other surgeons. It is alsodifficult to control how each respondent interpretsdifferent attributes and the language that is used todescribe each attribute. We attempted to minimizevariability between each respondent’s interpretationby providing a detailed description about each attributebefore asking the respondents to make a final choiceabout their preferred treatment. In addition, whendescribing the treatment options and attributes, theauthors purposefully used “plain English” to ensurethat the respondents adequately understood the optionspresented to them. We revised our language until thepilot survey demonstrated that participants understoodthe questions and descriptions well. However, the wayin which each scenario is described in a real-life clin-ical setting will undoubtedly differ from this surveyand will be framed differently by each provider. As

ol. 43, July 2018

678.e7 PATIENT PREFERENCES FOR SCAPHOID FRACTURES

such, the absolute percentages derived from this studydo not carry a specific meaning for individual patientsand surgeons. Future studies should use focus groupsand administer the survey in a clinical setting to buildupon the preliminary results of this study. Further-more, although conjoint analysis attempts to assignmonetary value to the trade-offs an individual iswilling to make for a form of treatment, this does notfully capture why a person may choose one treatmentover another. Although these results do provide aninitial perspective regarding a person’s treatmentpreferences, it does not reflect all the considerations aperson may contemplate in the decision-making pro-cess. Most notably, this survey does not consider theinfluence of postoperative pain, risks of postsurgicalcomplications (ie, infection), and the risks of anes-thesia, all of which may play a role in the decision-making process. Therefore, we cannot consider thisreport to offer a fully comprehensive perspective.

This study attempts to describe individual prefer-ences for the management of nondisplaced scaphoidfractures, and serves as a step toward the develop-ment of a tool to measure preferences about treatmentoptions in hand surgery. These initial results suggestthat people were more influenced by out-of-pocketcosts and their apprehension about undergoing sur-gery, than the time spent in a brace or cast, and therisk of their treatment failing. Future studies areneeded to further define the individual preferences ofpatients who experience this injury. The factors thatpatients deem important in their management maydiffer from the views of the provider, and the treat-ment should reflect the patient’s preferences.

REFERENCES

1. Shen L, Tang J, Luo C, Xie X, An Z, Zhang C. Comparison ofoperative and non-operative treatment of acute undisplaced orminimally-displaced scaphoid fractures: a meta-analysis of random-ized controlled trials. PLoS One. 2015;10(5):e0125247.

2. Puopolo SM, Rettig ME. Management of acute scaphoid fractures.Bull Hosp Jt Dis. 2003;61(3e4):160e163.

3. Adams JE, Steinmann SP. Acute scaphoid fractures. Orthop ClinNorth Am. 2007;38(2):229e235. vi.

4. Kawamura K, Chung KC. Treatment of scaphoid fractures andnonunions. J Hand Surg Am. 2008;33(6):988e997.

5. Vinnars B, Pietreanu M, Bodestedt A, Ekenstam F, Gerdin B.Nonoperative compared with operative treatment of acute scaphoidfractures. A randomized clinical trial. J Bone Joint Surg Am.2008;90(6):1176e1185.

6. Tong BC, Wallace S, Hartwig MG, D’Amico TA, Huber JC. Patientpreferences in treatment choices for early-stage lung cancer. AnnThorac Surg. 2016;102(6):1837e1844.

7. Tong BC, Huber JC, Ascheim DD, et al. Weighting compositeendpoints in clinical trials: essential evidence for the heart team. AnnThorac Surg. 2012;94(6):1908e1913.

8. Bernstein J. The variability of patient preferences. Clin Orthop RelatRes. 2012;470(7):1966e1972.

J Hand Surg Am. r V

9. Bunch WH, Chapman RG. Patient preferences in surgery for scoli-osis. J Bone Joint Surg Am. 1985;67(5):794e799.

10. Kocher MS, Bishop J, Marshall R, Briggs KK, Hawkins RJ. Oper-ative versus nonoperative management of acute Achilles tendonrupture: expected-value decision analysis. Am J Sports Med.2002;30(6):783e790.

11. Fraenkel L. Incorporating patients’ preferences into medical decisionmaking. Med Care Res Rev. 2013;70(1):80se93s.

12. Shackley P, Slack R, Michaels J. Vascular patients’ preferences forlocal treatment: an application of conjoint analysis. J Health Serv ResPolicy. 2001;6(3):151e157.

13. Ryan M, Farrar S. Using conjoint analysis to elicit preferences forhealth care. BMJ. 2000;320(7248):1530e1533.

14. Eberhart LH, Morin AM, Wulf H, Geldner G. Patient preferences forimmediate postoperative recovery. Br J Anaesth. 2002;89(5):760e761.

15. Maas A, Stalpers L. Assessing utilities by means of conjoint mea-surement: an application in medical decision analysis. Med DecisMaking. 1992;12(4):288e297.

16. Ryan M, Hughes J. Using conjoint analysis to assess women’spreferences for miscarriage management. Health Econ. 1997;6(3):261e273.

17. Vick S, Scott A. Agency in health care. Examining patients’ pref-erences for attributes of the doctor-patient relationship. J HealthEcon. 1998;17(5):587e605.

18. Hakim Z, Pathak DS. Modelling the EuroQol data: a comparison ofdiscrete choice conjoint and conditional preference modelling. HealthEcon. 1999;8(2):103e116.

19. O’Hara NN, Slobogean GP, Mohammadi T, et al. Arepatients willing to pay for total shoulder arthroplasty? Evidence froma discrete choice experiment. Can J Surg. 2016;59(2):107e112.

20. Johnson FR, Van Houtven G, Ozdemir S, et al. Multiple sclerosispatients’ benefit-risk preferences: serious adverse event risks versustreatment efficacy. J Neurol. 2009;256(4):554e562.

21. Phillips KA, Maddala T, Johnson FR. Measuring preferences forhealth care interventions using conjoint analysis: an application toHIV testing. Health Serv Res. 2002;37(6):1681e1705.

22. Waltzman JT, Scholz T, Evans GR. What patients look forwhen choosing a plastic surgeon: an assessment of patientpreference by conjoint analysis. Ann Plast Surg. 2011;66(6):643e647.

23. Kleinman L, McIntosh E, Ryan M, et al. Willingness to pay forcomplete symptom relief of gastroesophageal reflux disease. ArchIntern Med. 2002;162(12):1361e1366.

24. McIntosh E, Donaldson C, Ryan M. Recent advances in the methodsof cost-benefit analysis in healthcare. Pharmacoeconomics.1999;15(4):357e367.

25. Ratcliffe J. The use of conjoint analysis to elicit willingness-to-payvalues. Proceed with caution? Int J Technol Assess Health Care.2000;16(1):270e275.

26. Bala MV, Mauskopf JA, Wood LL. Willingness to pay as a measureof health benefits. Pharmacoeconomics. 1999;15(1):9e18.

27. Bond CD, Shin AY, McBride MT, Dao KD. Percutaneous screwfixation or cast immobilization for nondisplaced scaphoid fractures.J Bone Joint Surg Am. 2001;83-A(4):483e488.

28. Saeden B, Tornkvist H, Ponzer S, Hoglund M. Fracture of the carpalscaphoid. A prospective, randomised 12-year follow-up comparingoperative and conservative treatment. J Bone Joint Surg Br.2001;83(2):230e234.

29. Dias JJ, Dhukaram V, Abhinav A, Bhowal B, Wildin CJ. Clinical andradiological outcome of cast immobilisation versus surgical treatmentof acute scaphoid fractures at a mean follow-up of 93 months. J BoneJoint Surg Br. 2008;90(7):899e905.

30. Adolfsson L, Lindau T, Arner M. Acutrak screw fixation versus castimmobilisation for undisplaced scaphoid waist fractures. J Hand SurgBr. 2001;26(3):192e195.

31. McQueen MM, Gelbke MK, Wakefield A, Will EM, Gaebler C.Percutaneous screw fixation versus conservative treatment for

ol. 43, July 2018

PATIENT PREFERENCES FOR SCAPHOID FRACTURES 678.e8

fractures of the waist of the scaphoid: a prospective randomisedstudy. J Bone Joint Surg Br. 2008;90(1):66e71.

32. Dias JJ, Wildin CJ, Bhowal B, Thompson JR. Should acute scaphoidfractures be fixed? A randomized controlled trial. J Bone Joint SurgAm. 2005;87(10):2160e2168.

33. Davis EN, Chung KC, Kotsis SV, Lau FH, Vijan S. A cost/utilityanalysis of open reduction and internal fixation versus cast immo-bilization for acute nondisplaced mid-waist scaphoid fractures. PlastReconstr Surg. 2006;117(4):1223e1235; discussion 1236-1228.

34. Beusterien K, Grinspan J, Kuchuk I, et al. Use of conjoint analysis toassess breast cancer patient preferences for chemotherapy side ef-fects. Oncologist. 2014;19(2):127e134.

35. Johansson G, Ställberg B, Tornling G, et al. Asthma treatmentpreference study: a conjoint analysis of preferred drug treatments.Chest. 2004;125(3):916e923.

36. Bridges JFP, Mohamed AF, Finnern HW, Woehl A, Hauber AB.Patients’ preferences for treatment outcomes for advanced non-smallcell lung cancer: a conjoint analysis. Lung Cancer. 2012;77(1):224e231.

37. Amazon Web Services. Introduction to Amazon Mechanical Turk.Available at: http://docs.aws.amazon.com/AWSMechTurk/latest/AWSMechanicalTurkGettingStartedGuide/SvcIntro.html. AccessedJanuary 16, 2017.

38. Buhrmester M, Kwang T, Gosling SD. Amazon’s Mechanical Turk: anew source of inexpensive, yet high-quality, data? Perspect PsycholSci. 2011;6(1):3e5.

39. Shapiro DN, Chandler J, Mueller PA. Using Mechanical Turk tostudy clinical populations. Clin Psychol Sci. 2013;1(2):213e220.

40. Berinsky AJ, Huber GA, Lenz GS. Evaluating online labor marketsfor experimental research: Amazon.com’s Mechanical Turk. PoliticalAnalysis. 2012;20(3):351e368.

J Hand Surg Am. r V

41. AristidesM,Chen J, SchulzM,WilliamsonE,ClarkeS,GrantK.Conjointanalysis of a newChemotherapy: willingness to pay and preference for thefeatures of raltitrexed versus standard therapy in advanced ColorectalCancer. Pharmacoeconomics. 2002;20(11):775e784.

42. Kim WL, Kim JS, Lee JB, Kim SH, Min DU, Park HY. Survey ofpreferences in patients scheduled for carpal tunnel release usingconjoint analysis. Clin Orthop Surg. 2017;9(1):96e100.

43. Streufert B, Reed SD, Orlando LA, Taylor DC, Huber JC,Mather RC III. Understanding preferences for treatment after hypo-thetical first-time anterior shoulder dislocation: surveying an onlinepanel utilizing a novel shared decision-making tool. Orthop J SportsMed. 2017;5(3). 2325967117695788.

44. Sawtooth Software. The ACA/WEB v.6.0 Technical Paper. North-brook, IL: Sawtooth Software; 2007.

45. Allenby G, Fennell G, Huber J, et al. Adjusting choice models tobetter predict market behavior. Mark Lett. 2005;16(3):197e208.

46. Hampson LA, Allen IE, Gaither TW, et al. Patient-centered treatmentdecisions for urethral stricture: conjoint analysis improves surgicaldecision-making. Urology. 2017;99:246e253.

47. Butler GS, Hurley CA, Buchanan KL, Smith-VanHorne J. Pre-hospital education: effectiveness with total hip replacement surgerypatients. Patient Educ Couns. 1996;29(2):189e197.

48. Gammon J, Mulholland CW. Effect of preparatory information priorto elective total hip replacement on psychological coping outcomes.J Adv Nurs. 1996;24(2):303e308.

49. Daltroy LH, Morlino CI, Eaton HM, Poss R, Liang MH. Preoperativeeducation for total hip and knee replacement patients. Arthritis CareRes. 1998;11(6):469e478.

50. Wallace LM. Pre-operative state anxiety as a mediator of psycho-logical adjustment to and recovery from surgery. Br J Med Psychol.1986;59(Pt 3):253e261.

ol. 43, July 2018

TABLE E1. Evidence Used to Formulate Attributes and Levels for the Conjoint Analysis

AuthorYear

Length of Immobilization(Cast/Brace)

Length of RemainingSoreness and/or Stiffness

Risk of ScaphoidNonunion

Cost(Direct Costs)

Saeden et al200128

C: 12 � 3 wkS: 2 � 1 wk

Not assessed C: 7%S: 3%

e

Bond et al200127

C: 6 wk—union(average of 12 wk)

S: until union(average of 7 wk)

Not assessed C: 0%S: 0%

e

Adolfsson et al200130

C: 10 wkS: 3 wk

Not assessed C: 0%S: 4%

e

Dias et al200532 and 200829

C: 8 wkS: no immobilization

Pain*:C: 0/2/6 mo: (4.45/2.29/2.35)S: 0/2/6 mo: (4.05/2.39/2.40)Tenderness†:C: 0/2/6 mo: (2.21/0.76/0.17)S: 0/2/6 mo: (2.14/0.55/0.26)

C: 23%S: 0%

e

McQueen et al200831

C: 8e12 wkS: no immobilization

Not assessed C: 13%S: 3%

e

Vinnars et al20085

C: mean of 10 wkS: mean of 3 wk

Not assessed C: 2%S: 0%

e

Davis et al200633

e e e C: $605(25e64 y)

S: $1,747(25e64 y)

Levels includedin conjoint analysis

Cast: 2, 4, 8 wkfollowed byBrace: 2, 4, 8 wk

0 mo2 mo6 mo

3%5%10%

$500$1,000$2,500

C, cast immobilization; S, surgery.*Pain was assessed on a scale of 1e10 using the visual analog scale pain score.†Tenderness was assessed on a scale of 1e7 using the Patient Evaluation Measure Questionnaire 10.

678.e9 PATIENT PREFERENCES FOR SCAPHOID FRACTURES

J Hand Surg Am. r Vol. 43, July 2018