SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Objective

To examine the impact of a videobooklet patient decision aid supplemented by an interactive values clarification exercise on decisional conflict in patients with knee osteoarthritis (OA) considering total knee arthroplasty.

Methods

A total of 208 patients participated in the study (mean age 63 years, 68% female, and 66% white). Participants were randomized to 1 of 3 groups: 1) educational booklet on OA management (control), 2) patient decision aid (videobooklet) on OA management, and 3) patient decision aid (videobooklet) + adaptive conjoint analysis (ACA) tool. The ACA tool enables patients to consider competing attributes (i.e., specific risks/benefits) by asking them to rate a series of paired comparisons. The primary outcome was the decisional conflict scale ranging from 0–100. Differences between groups were analyzed using analysis of variance and Tukey's honestly significant difference tests.

Results

Overall, decisional conflict decreased significantly in all groups (P < 0.05). The largest reduction in decisional conflict was observed for participants in the videobooklet decision aid group (21 points). Statistically significant differences in pre- versus postintervention total scores favored the videobooklet group compared to the control group (21 versus 10) and to the videobooklet plus ACA group (21 versus 14; P < 0.001). Changes in the decisional conflict score for the control group compared to the videobooklet decision aid + ACA group were not significantly different.

Conclusion

In our study, an audiovisual patient decision aid decreased decisional conflict more than printed material alone or the addition of a more complex computer-based ACA tool requiring more intense cognitive involvement and explicit value choices.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Current treatments for knee osteoarthritis (OA) alleviate symptoms, but do not affect disease progression. For patients with pain and disability unresponsive to medical treatment, total knee arthroplasty (TKA) results in substantial pain relief and functional improvement (1). In 2006, the rate of TKA in the US was more than double that of the previous decade (2) and is expected to continue to rise with advanced aging of the population (3).

Despite its benefits, TKA is not without risks. Perioperative complications can occur and, furthermore, prosthesis failure may require surgical revision. Because TKA is an elective procedure and patients must carefully weigh potential benefits and risks, the decision-making process is complex and uncertainty is common. Decisions are strongly related to patients' preexisting preferences and values and the patient-physician relationship (4, 5). In addition, some groups such as African Americans and older patients are less likely to choose surgery over conservative therapy (6–10). Qualitative research has demonstrated that patient misperceptions of surgery in general, misunderstanding of the etiology of OA and its progressive nature, low expectations of TKA outcomes, and fear of surgical complications all influence patient preferences for this procedure (5, 11–13).

Clear communication between the patient and physician leading to mutual understanding of the risks and benefits of treatment options for OA increases patient confidence and may lead to higher satisfaction with a decision (13). However, the clinical encounter is usually short, and often there is not sufficient time to discuss thoroughly the complex issues that patients must understand to clarify their values and make a decision. Patient decision aids can reduce the level of uncertainty or “decisional conflict” (14). These tools increase patients' knowledge about the risks and benefits of therapeutic alternatives, help them clarify their values and preferences, and prepare them for the encounter with their physician and deciding on a course of action (14, 15).

Conjoint analysis has been used for over a decade to explore patient preferences in health care (16). This approach enables patients to evaluate tradeoffs between 2 or more scenarios that present different levels of attributes (risks and benefits) related to the health or therapeutic problem presented to them. Adaptive conjoint analysis (ACA) is a specific type of conjoint analysis that has been explored extensively by Fraenkel et al to describe patient preferences and values impacting decision making in rheumatic conditions (17–19). ACA collects and analyzes preference data using an interactive computer program (Sawtooth Software). ACA is unique in that it uses an individual respondent's answers to update and refine questions through a series of paired comparisons (Figure 1). In addition, because it forces respondents to cognitively engage in the choice between alternatives with competing risks and benefits, it is considered a values clarification exercise of the various attributes pertinent to the decision. The assumption is that the interactive exercise provides the patient with a “deeper insight into his or her own constellation of decision-relevant values” (20).

thumbnail image

Figure 1. Screenshot of adaptive conjoint analysis program.

Download figure to PowerPoint

We conducted a prospective, 3-arm randomized controlled trial (RCT) to compare ACA coupled with a videobooklet patient decision aid to the videobooklet alone and to standard educational materials regarding patients' decisional conflict associated with TKA for knee OA.

Significance & Innovations

  • Patient decision aids provide information to patients regarding risks and benefits of multiple treatment options and have been shown to reduce decisional conflict.

  • Reduction in decisional conflict or uncertainty is linked to improved satisfaction with the decision.

  • Comprehensive decision aids appear to reduce decisional conflict in osteoarthritis patients considering total knee arthroplasty and could be used at point of care to facilitate informed patient decision making.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Participant recruitment.

Participants were recruited using multiple methods, including advertisements in several local newspapers, on Facebook, and by contacting participants of a previous research study. Participants were considered eligible if: 1) a physician had told them they had OA of the knee diagnosed with a radiograph at least 2 years before screening, 2) their knee OA interfered with their activities of daily living, 3) they experienced pain (at least a 4 on a scale of 1–10) on most days in the last 3 months, and 4) they had ever considered or talked to a doctor about TKA. Patients were excluded if they had rheumatoid arthritis, had not had radiographs of their knees, had undergone TKA or were currently scheduled for TKA, if they were not comfortable reading and communicating in English, or if they were not comfortable answering questions on a computer using a mouse. After providing informed consent, patients were scheduled to participate in the trial. The study was approved by the Institutional Review Board of the University of Texas MD Anderson Cancer Center and Baylor College of Medicine.

ACA program.

Eight attributes were selected via consensus by 2 rheumatologists (LF and MES-A) to evaluate patient preferences for TKA surgery. Characteristics of attributes associated with TKA were then reviewed in the literature to determine 4 levels for each characteristic chosen (Table 1). Characteristics of knee pain and function were derived from items from the Western Ontario and McMaster Universities Osteoarthritis Index, specifically “having little or no pain when you walk,” “having little or no pain at night,” and “being able to do activities like shopping or golfing with little or no difficulty.” Due to recent research regarding the presentation of risk perception (21–23), we utilized frequency format to describe risk and added verbal descriptions generating a combined frame presented in the Introduction and continued throughout the program. In addition, we utilized visual pictographs to enable understanding of risk and generate affective response.

Table 1. Attributes and levels utilized for adaptive conjoint analysis program
AttributeLevel 1Level 2Level 3Level 4
Risk of serious complication (lung clot, pneumonia, or heart attack)No increased risk1 in 1005 in 10010 in 100
Extremely rare risk of death from surgeryNo increased risk1 in 1,0005 in 1,00010 in 1,000
Need for another operation after 15 yearsNo need2 in 1008 in 10015 in 100
No. days in hospitalNo days in hospital3 days5 days15 days
Need for rehabilitation and walking aidsNo need4 weeks6 weeks12 weeks
Having little/no pain when you walkSame amount of pain5 out of 107 out of 109 out of 10
Having little/no pain at nightSame amount of pain5 out of 107 out of 109 out of 10
Able to do activities (shopping/golfing) with little/no difficulty 6 months from nowSame amount of difficulty5 out of 107 out of 109 out of 10

Patients were first presented with a list of the 8 characteristics and were told to choose only 1 as being the most important thing to them “when thinking about knee replacement surgery.” They then were led to a screen where they ranked each of the 7 other attributes on a scale from 1–10 (“not nearly as important” to “just as important”) compared to the first chosen attribute (18). These rankings were used by the software to generate an initial estimate of each subject's values. Next, participants were presented with 2 treatment choices (described to be “exactly the same except for the differences [provided]”) and asked to select their preference (weighted scale ranging from “strongly prefer left” to “strongly prefer right”) (Figure 1). Each treatment option provided was described with the same 2 attributes at 2 different levels. Lastly, participants were presented with results of the program's calculation of how important each fact was to them. These were presented with bar graphs, the longer bars representing increased importance (Figure 2). These results were explained by the research assistant, printed, and given to the participants.

thumbnail image

Figure 2. Example of adaptive conjoint analysis output.

Download figure to PowerPoint

We also examined the participants' preferred treatment before and after the intervention. We generated a dichotomized variable indicating certainty in treatment choice (i.e., chose preferred treatment as “knee replacement surgery” or “pills and physical therapy” versus “unsure”).

Procedures.

Study participants were randomized into 1 of the 3 intervention arms using a computer-generated random list with unequal blocks and numbered, sealed, and opaque envelopes to ensure allocation concealment. Patients were randomized after completing the baseline assessment and only then was the assigned envelope opened to allocate patient to group.

Subjects randomized to group 1 (control) were given a printed booklet about treatment choices for knee OA, including medical management and surgery, published by the National Institute of Arthritis and Musculoskeletal and Skin Diseases (24). Participants were asked to read the booklet, which took ∼20 minutes.

Subjects randomized to group 2 (videobooklet) were given a videobooklet decision aid developed by Health Dialog with the Foundation for Informed Medical Decision-Making (FIMDM) entitled “Treatment Choices for Knee Osteoarthritis,” including a DVD and a booklet to follow along while viewing the DVD. The Shared Decision-Making video was ∼45 minutes long and met criteria created by the International Patient Decision Aids Standards Collaboration in the areas of content, development process, and effectiveness (25).

Subjects randomized to group 3 (videobooklet+ACA) were given the same videobooklet decision aid to watch and read first. Afterward, they received a brief training session regarding the computer-based ACA tool and then completed the ACA tool, which took an additional 15 minutes.

On the day of the scheduled intervention and after randomization, participants assigned to the control group were taken to a room where they were provided with the educational book and, upon completion of the reading, they completed the postintervention questionnaires. Participants in the videobooklet and videobooklet+ACA groups remained together and viewed the video. Once the video was finished, the groups were again divided and the videobooklet group was taken to a separate room where they completed their postintervention questionnaires. The videobooklet+ACA group was given instruction on the ACA program by the study coordinator (SdA). A demonstration ACA program (using the same instructions and introduction screens as the program used by the individual patients) was presented to the participants while the coordinator explained the exercise. Participants were allowed to ask questions about the program during the demonstration. Upon completion of the demonstration, participants were led to individual computers, where they completed the ACA program on their own. The study coordinator was available to answer questions about use. After participants completed the program, they were given postintervention questionnaires to complete.

Measures.

Primary outcome measure.

The decisional conflict scale developed by O'Connor was adapted for knee OA to evaluate the participants' uncertainty with their treatment options (15). Patients were first asked which treatment option they preferred, i.e., “knee replacement surgery; pills and physical therapy, possibly occasional knee joint injections; or unsure.” The total score ranges from 0 (no decisional conflict) to 100 (extremely high decisional conflict). The decisional conflict scale has 5 subscales (all 0–100): 1) uncertainty, 2) informed, 3) values clarity, 4) support, and 5) effective decision. For all subscales, higher scores indicate greater decisional conflict. The decisional conflict scale was administered before the intervention and immediately after. O'Connor has generated thresholds of meaningful differences (15). A score of <25 out of 100 is “associated with implementing decisions,” and a score of >37.5 out of 100 is “associated with decision delay or feeling unsure about implementation.” The total score was used as the primary outcome measure and the subscale scores were used as secondary outcome measures (26).

Covariates.

Demographic information collected included self-reported age, race/ethnicity, sex, education, marital status, and employment status. Comorbidities were assessed using participants' self-reported checklist of conditions generated from the conditions present in the Charlson Comorbidity Index (27). Body mass index was calculated from self-reported height and weight. The Knee Injury and Osteoarthritis Outcome Score (KOOS) was utilized to measure knee pain and symptoms and the effects of arthritis on participants' activities of daily living, sports and recreational activities, and quality of life (28). Patients also indicated in which knee they had OA (right, left, or both) and the year in which they were first diagnosed as having OA.

Statistical analysis.

All analyses were conducted according to group allocation. A sample size of 70 per group was calculated to be sufficient to estimate a moderate effect size (0.5) responsive to change in the decisional conflict scale before and after interventions (26) with a 2-tailed probability of a Type I error of 0.05, and a Type II error of 0.20 (80% power). Chi-square test and analysis of variance were utilized to identify differences in patient characteristics for categorical and continuous variables, respectively. We defined change in decisional conflict total score and subscale scores as the difference between the postintervention and preintervention scores (so that a negative difference indicated a reduction in decisional conflict). Tukey's honestly significant difference post hoc test was utilized to examine differences between randomized groups. Statistically significant differences were defined as P values less than 0.05. We also conducted analysis of covariance (ANCOVA) to control for potential baseline differences among groups. Decisional conflict scale postintervention scores were the dependent variable in the ANCOVA model, with group as the independent variable and baseline scores as a covariate. All analyses were conducted using Stata statistical software, version 10 (StataCorp).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Patient disposition.

A total of 1,134 potential participants were contacted by study staff; 64% were deemed ineligible after screening or were unable to be contacted after multiple attempts before the end of the recruitment period. Of the remaining eligible participants, 31% declined participation and 18% were unable to attend the intervention session before the end of the recruitment period. A total of 210 participants were randomized into 1 of the 3 treatment groups. Two participants (1 from the control group and 1 from the videobooklet+ACA group) did not turn in the postintervention questionnaire and were excluded in this analysis, leaving a total of 208 participants in the analysis (Figure 3).

thumbnail image

Figure 3. A Consolidated Standards of Reporting Trials diagram of participants. DCS = decisional conflict scale.

Download figure to PowerPoint

Patient characteristics.

Patient characteristics are described in Table 2. After randomization, no statistically significant differences existed among groups in participant demographic characteristics, KOOS pain, function, or quality of life scores. Statistically significant differences (P = 0.02), however, were noted between the videobooklet group and the videobooklet+ACA group in KOOS symptom scores, but not on the other KOOS subscales. The difference was small (5 points) with those in the videobooklet group reporting less knee symptoms.

Table 2. Participant characteristics*
 Total (n = 208)Group 1 control (n = 69)Group 2 videobooklet (n = 70)Group 3 videobooklet + ACA (n = 69)
  • *

    Values are the number (percentage) unless indicated otherwise. ACA = adaptive conjoint analysis; KOOS = Knee Injury and Osteoarthritis Outcome Score.

  • P < 0.05 by analysis of variance. Group 2 significantly higher than group 3 using Tukey's honestly significant difference. No other statistically significant differences by group.

Women141 (68)48 (70)47 (67)46 (67)
Age, mean ± SD years62.8 ± 9.063.6 ± 8.662.6 ± 9.262.5 ± 9.4
White137 (66)49 (71)49 (70)39 (56)
African American49 (24)14 (20)17 (24)18 (26)
Hispanic15 (7)3 (4)2 (3)10 (14)
Other7 (3)3 (4)2 (3)2 (3)
Less than undergraduate degree91 (44)30 (43)28 (40)33 (48)
Undergraduate/advanced degree117 (56)39 (56)42 (60)36 (52)
Disabled13 (6)3 (4)6 (8)4 (6)
Employed76 (36)25 (36)24 (34)27 (39)
Retired85 (41)29 (42)28 (40)28 (40)
Other than employed/retired34 (16)12 (17)12 (17)10 (14)
Bilateral knee symptoms116 (56)42 (61)38 (54)36 (52)
Disease duration, mean ± SD years9.6 ± 8.210.6 ± 9.89.0 ± 6.69.2 ± 7.9
Body mass index, mean ± SD kg/m232.9 ± 8.132.0 ± 7.934.3 ± 8.732.3 ± 7.7
Comorbidity index, mean ± SD0.9 ± 1.60.8 ± 1.61.1 ± 1.80.7 ± 1.3
KOOS pain, mean ± SD48.8 ± 17.550.0 ± 18.450.7 ± 15.245.7 ± 18.6
KOOS symptom, mean ± SD42.4 ± 11.241.6 ± 10.845.3 ± 11.540.3 ± 10.7
KOOS function in daily living, mean ± SD51.9 ± 19.052.5 ± 17.754.4 ± 18.348.8 ± 20.6
KOOS function in sports/recreation, mean ± SD25.5 ± 24.926.4 ± 25.526.4 ± 25.323.8 ± 24.1
KOOS quality of life, mean ± SD27.5 ± 17.729.2 ± 17.528.2 ± 16.625.0 ± 18.9

Pre-post results after intervention.

Prior to the intervention, a statistically significant difference was noted between the control and videobooklet groups in the effective decision subscale (Table 3). No statistically significant differences between groups were noted for the total score or the other subscales at baseline. Immediately after the intervention, the decisional conflict scale total score and subscale scores were significantly lower in all 3 groups, indicating a reduction in decisional conflict. Post hoc analyses of differences between groups indicated that both the videobooklet and videobooklet+ACA interventions had greater reductions in decisional conflict than the booklet alone (−21.0, −14.0, and −9.5, respectively; P < 0.001). In addition, the reduction in decisional conflict was greater for the videobooklet group than the videobooklet+ACA group.

Table 3. Multiple comparison group differences of DCS total and subscale scores*
DCS variableGroup 1 control (n = 69)Group 2 videobooklet (n = 70)Group 3 videobooklet + ACA (n = 69)PStatistically significant comparisons
  • *

    Values are the mean ± SEM unless indicated otherwise. DCS = decisional conflict scale; ACA = adaptive conjoint analysis.

  • Tukey's honestly significant difference test, P < 0.05 (analysis of variance [ANOVA]).

  • Overall and within groups, postintervention scores statistically significantly differed from preintervention scores, P < 0.05 (ANOVA).

Total score pre38.8 ± 2.242.7 ± 2.337.4 ± 2.0> 0.20None
Total score post29.2 ± 2.021.6 ± 1.523.4 ± 1.80.0091 vs. 2 and 1 vs. 3
Total score difference (pre-post)−9.5 ± 1.8−21.0 ± 2.2−14.0 ± 1.9< 0.0011 vs. 2 and 2 vs. 3
Uncertainty pre46.4 ± 3.152.5 ± 3.048.1 ± 2.9> 0.20None
Uncertainty post35.9 ± 2.729.8 ± 2.233.4 ± 2.8> 0.20None
Uncertainty difference (pre-post)−10.5 ± 2.4−22.7 ± 3.2−14.6 ± 3.00.011 vs. 2
Informed pre41.4 ± 2.942.6 ± 2.837.1 ± 2.6> 0.20None
Informed post27.3 ± 2.015.4 ± 1.715.9 ± 1.9< 0.0011 vs. 2 and 1 vs. 3
Informed difference (pre-post)−14.1 ± 2.6−27.3 ± 2.7−21.1 ± 2.50.0021 vs. 2
Values clarity pre38.5 ± 2.937.4 ± 2.532.2 ± 2.6> 0.20None
Values clarity post26.1 ± 2.316.9 ± 1.817.9 ± 1.80.0021 vs. 2 and 1 vs. 3
Values clarity difference (pre-post)−12.4 ± 2.6−20.5 ± 2.7−14.4 ± 2.40.071 vs. 2
Support pre32.5 ± 2.136.0 ± 2.432.8 ± 2.0> 0.20None
Support post25.0 ± 1.919.9 ± 1.720.5 ± 1.80.10None
Support difference (pre-post)−7.5 ± 2.1−16.1 ± 2.3−12.3 ± 1.90.021 vs. 2
Effective decision pre36.0 ± 2.444.4 ± 2.537.0 ± 2.10.021 vs. 2
Effective decision post31.2 ± 2.325.1 ± 1.927.7 ± 2.20.12None
Effective decision difference (post-pre)−4.7 ± 1.8−19.3 ± 2.6−9.3 ± 2.5< 0.0011 vs. 2 and 2 vs. 3

The overall baseline decisional conflict total score was ∼40, indicating high decisional conflict and uncertainty. Results on the decisional conflict subscales demonstrated similar patterns, with both videobooklet groups showing significant reductions in decisional conflict compared to controls. The largest reduction was observed in the videobooklet alone group. Overall, the 2 decisional conflict subscales that improved the most were “informed” and “uncertainty.” The mean difference in pre- and postintervention total and subscale decisional conflict scores varied from −4.5 to −27.3 out of 100. After the intervention, the overall decisional conflict total group mean score was slightly less than 25, suggesting patients were ready to implement a decision about treatment.

In the ANCOVA model, preintervention decisional conflict scores and randomized group were significantly associated with postintervention decisional conflict total score. Due to statistically significant differences between groups in the KOOS symptom scores at baseline, an additional ANCOVA model was generated, taking into account the patients' self-reported symptoms measured by the KOOS. When added to the model, KOOS symptoms did not significantly contribute to a change in the decisional conflict total or subscale scores.

Among the 208 patients who indicated a specific preferred treatment before the intervention (either “knee replacement surgery” or “pills and physical therapy, possibly occasional knee joint injections”), 87.5% continued to indicate a preferred choice after the intervention, while 12.5% became “unsure” about their preferred treatment. Among the 80 patients who were initially “unsure” about their preferred treatment, 61.2% chose a treatment option while 38.8% remained “unsure.” We also examined if change in patient preferences about OA treatment could have an impact on decisional conflict after receipt of the decision aids. In order to further evaluate this possibility, we incorporated a dichotomized variable indicating change in preference in the pre- versus postintervention decisional conflict scale total score into an ANCOVA model. The analysis demonstrated the change in preference itself was not statistically associated with decisional conflict scores.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Patients with advanced knee OA are often faced with the decision about TKA. This decision can be difficult, particularly if patients do not clearly understand the benefits and risks of each option (undergoing surgery or continuing medical management alone). It is important to ensure that patients are effectively informed and decisions are not guided by misperceptions and false beliefs. The decision process itself involves multiple health care providers and discussions regarding pros and cons of surgery and patient preferences (29).

In this study, we examined the effect of a decision aid supplemented by an ACA program on decisional conflict regarding treatment options in individuals with knee OA. We used 3 different tools: 1) an educational booklet developed by the National Institutes of Health, 2) a videobooklet decision aid developed by FIMDM, and 3) a computer-based ACA values clarification exercise with the videobooklet decision aid. Decisional conflict decreased among all patient groups receiving information about treatment options for OA. This finding was expected and supports theory in which information provided to patients about risks and benefits of treatment options decreases decisional conflict. Previous studies measuring change in the decisional conflict after decision aids range from −2.5 to −17.1 (14).

The magnitude of reduced decisional conflict in our cohort varied between groups. Those who received the control brochure had the smallest reduction in decisional conflict, while those who viewed the videobooklet had the greatest reduction in decisional conflict. Furthermore, those in the control group continued to have scores greater than 25, indicating the potential for high decisional conflict and uncertainty, while those in the videobooklet and videobooklet+ACA group had scores below 25, indicating readiness to implement a decision. Both groups receiving the videobooklet intervention had increased improvement with respect to individuals receiving the booklet alone. The videobooklet decision aid not only described OA and the potential treatments like the control group, but also included outcomes of the options; communicated risks, benefits, and uncertainties using audiovisual and graphic methods; and suggested the patients take their own values into account when deciding about TKA. In addition, the videobooklet integrated testimonials of patients that had chosen surgical and nonsurgical treatment options and emphasized communication with the patient's physician, which may have added to its overall effectiveness when compared to the booklet (30–32). The various components of the decision aid provide a possible explanation to the greater magnitude in reduced decisional conflict found in our study compared to previous studies. Weng et al utilized the same videobooklet produced by FIMDM to determine the effect of the decision aid on knowledge, expectations, and decisional conflict (33). They found the video to be effective in creating realistic expectations. In addition, viewers of the video felt more confident in their knowledge of the efficacy of TKA treatment. In their cohort, the postintervention decisional conflict score was 25.8, which is similar to our cohorts' final score of 25 (33).

Previous RCTs have compared patient decision aids to usual care and simple to complex decision aids. To our knowledge, few have been utilized in the area of decision making for OA treatment, utilizing TKA as a treatment option. As summarized by O'Connor et al, 10 studies compared decision aids to usual care for a variety of decisions and evaluated decisional conflict (14). Decisional conflict decreased in all of the included studies when comparing the decision aid versus usual care. In the meta-analysis, O'Connor and colleagues also included a comparison of the impact of simple versus complex decision aids on decisional conflict. Their results indicated only 1 of 7 studies demonstrated significant reductions in decisional conflict with detailed versus simple decision aids. This is comparable to our results in that we did not see a further decrease in decisional conflict among the videobooklet+ACA group when compared to the group that received the videobooklet only (14, 34).

Potential respondent fatigue could be one explanation to our findings: participants in the videobooklet group took approximately 1 hour to view the video and complete the questionnaires, while those in the videobooklet+ACA took on average 2 hours. To further investigate the potential effect of respondent fatigue on decisional conflict, we evaluated scale reliability of the total decisional conflict post scores among the 3 groups, using Cronbach's alpha and mean to variance ratios to determine if for those engaged in the videobooklet+ACA intervention variability increased (and reliability decreased) in their responses. The internal consistency was excellent and consistent between groups (0.96, 0.94, and 0.94 for control, videobooklet, and videobooklet+ACA groups, respectively). Mean to variance ratios were 9.71, 7.76, and 9.60 for control, videobooklet, and videobooklet+ACA groups, respectively. No statistically significant differences were observed in the mean to variance ratio between groups. This suggests that respondent fatigue did not have an impact on the reliability of the outcome measure.

Another potential explanation for the greater reduction in decisional conflict among the videobooklet only group is that the ACA exercise, in addition to the information provided by the videobooklet, was cognitively rigorous and in fact reduced clarity about the decision, increasing uncertainty. It has been suggested that complex decisions do not always benefit from intense cognitive evaluation, and that in these circumstances “gist” driven decisions may be more satisfying (35). We only assessed ACA in addition to the videobooklet decision aid. Previous studies in which patient preferences were elicited utilizing conjoint analysis alone demonstrated that patients are able to evaluate multiple risks, benefits, and uncertainties regarding specific treatment options (19). Fraenkel et al utilized ACA to elicit preferences regarding treatment options for OA including nonsteroidal antiinflammatory drugs, cyclooxygenase 2 inhibitors, glucosamine and/or chondroitin, opioids, or capsaicin (19). They found patients preferred topical capsaicin most likely due to the low probability of adverse effects. However, patients in this study did not receive additional information about the decision from another decision aid. Fraenkel et al also did not include TKA as a potential treatment option. In addition, decisional conflict was not evaluated.

Although the content of the decision aids utilized in this study was comprehensive, additional specific content areas were not included. The aids utilized in this study did not include content regarding costs of the treatments. This may be an area of concern for many with OA, considering the cost of TKA can be more than $100,000 for a person at end-stage OA and with multiple comorbidities (36). Also, it is important to note that the control group used in this study did not receive “standard of care.” Rather, the control group received publically available educational materials about OA and treatment options, offering a reasonable comparison with the 2 decision aid groups.

Our participants were age >55 years, mostly female, and with a high prevalence of obesity, characteristics that are representative of patients with knee OA at large (37). However, they were mostly white and educated and may not be representative of OA patients from ethnically diverse backgrounds or low literacy levels considering treatment options. Our recruitment methods may have generated volunteer bias. Less than 20% of the participants were recruited from the previous study and over 65% of our participants were recruited from newspaper ads. Recruitment from the waiting room of a physician visit may have generated a different population; however, we feel our recruitment methods generated a more representative group of all patients with knee OA, not only those who were receiving current care for their OA.

Participants described experiences of friends or family members who had already undergone TKA. The experiences of others may have influenced the decision-making process and affected it toward surgical options; however, we did not formally evaluate these influences. Future evaluations of treatment decision making should incorporate experiences and heuristics.

The ACA program may have increased decisional conflict after the delivery of the videobooklet. However, this may not be a negative effect of the ACA program. It is important to consider the variation within our participant group in the stage of decision regarding TKA. Although decisional conflict is generally assessed as a negative attribute of a decision, it is useful to think that decisional conflict may be expected for many in this population and they work through the decision. Those that viewed the videobooklet and participated in the ACA program were able to view the risks and the benefits of TKA specifically and may have delayed their decision or become uncertain regarding their preferences. This does not necessarily indicate that they will not make a quality decision or be satisfied with the decision. Instead, it suggests additional thought into the decision.

Our findings suggest that a comprehensive videobooklet including a decision aid significantly improved decisional conflict in patients with knee OA when considering multiple treatment options. Participating in the ACA exercise, however, did not contribute to additional reductions in decisional conflict. In addition, certainty about treatment decisions can vary after exposure to information, suggesting that early patient education about treatment options for OA during the discussion with the primary care physician or rheumatologist is needed. Long-term effectiveness of these tools on the quality of the decision, overall satisfaction with the decision, and final impact on preferences about TKA is yet to be determined.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Suarez-Almazor 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.

Study conception and design. De Achaval, Fraenkel, Volk, Suarez-Almazor.

Acquisition of data. De Achaval, Suarez-Almazor.

Analysis and interpretation of data. De Achaval, Fraenkel, Volk, Cox, Suarez-Almazor.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

The authors are grateful to Mr. Vincent Richards who assisted us in the recruitment and followup of participants.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES
  • 1
    Jones CA, Beaupre LA, Johnston DW, Suarez-Almazor ME. Total joint arthroplasties: current concepts of patient outcomes after surgery. Rheum Dis Clin North Am 2007; 33: 7186.
  • 2
    Freid VM, Bernstein AB. Health care utilization among adults aged 55-64 years: how has it changed over the past 10 years? NCHS Data Brief 2010; 32: 18.
  • 3
    Singh JA, Vessely MB, Harmsen WS, Schleck CD, Melton LJ 3rd, Kurland RL, et al. A population-based study of trends in the use of total hip and total knee arthroplasty, 1969-2008. Mayo Clin Proc 2010; 85: 898904.
  • 4
    Katz JN, Lyons N, Wolff LS, Emrani P, Holt HL, Corbett KL, et al. Medical decision-making among Hispanics and non-Hispanic whites with chronic back and knee pain: a qualitative study. BMC Musculoskelet Disord 2011; 12: 78.
  • 5
    O'Neill T, Jinks C, Ong BN. Decision-making regarding total knee replacement surgery: a qualitative meta-synthesis. BMC Health Serv Res 2007; 7: 52.
  • 6
    Kroll TL, Richardson M, Sharf BF, Suarez-Almazor ME. “Keep on truckin'” or “It's got you in this little vacuum”: race-based perceptions in decision-making for total knee arthroplasty. J Rheumatol 2007; 34: 106975.
  • 7
    Suarez-Almazor ME, Souchek J, Kelly PA, O'Malley K, Byrne M, Richardson M, et al. Ethnic variation in knee replacement: patient preferences or uninformed disparity? Arch Intern Med 2005; 165: 111724.
  • 8
    Byrne MM, Souchek J, Richardson M, Suarez-Almazor M. Racial/ethnic differences in preferences for total knee replacement surgery. J Clinic Epidemiol 2006; 59: 107886.
  • 9
    Figaro MK, Russo PW, Allegrante JP. Preferences for arthritis care among urban African Americans: “I don't want to be cut.” Health Psychol 2004; 23: 3249.
  • 10
    Hamel MB, Toth M, Legedza A, Rosen MP. Joint replacement surgery in elderly patients with severe osteoarthritis of the hip or knee: decision making, postoperative recovery, and clinical outcomes. Arch Intern Med 2008; 168: 143040.
  • 11
    Wright JG, Santaguida PL, Young N, Hawker GA, Schemitsch E, Owen JL. Patient preferences before and after total knee arthroplasty. J Clin Epidemiol 2010; 63: 77482.
  • 12
    Suarez-Almazor ME, Richardson M, Kroll TL, Sharf BF. A qualitative analysis of decision-making for total knee replacement in patients with osteoarthritis. J Clin Rheumatol 2010; 16: 15863.
  • 13
    Street RL Jr, Richardson MN, Cox V, Suarez-Almazor ME. (Mis)understanding in patient-health care provider communication about total knee replacement. Arthritis Rheum 2009; 61: 1007.
  • 14
    O'Connor AM, Bennett CL, Stacey D, Barry M, Col NF, Eden KB, et al. Decision aids for people facing health treatment or screening decisions. Cochrane Database Syst Rev 2009; 3:CD001431.
  • 15
    O'Connor AM. Validation of a decisional conflict scale. Med Decis Making 1995; 15: 2530.
  • 16
    Marshall D, Bridges JF, Hauber B, Cameron, R, Donnalley L, Fyie K, et al. Conjoint analysis applications in health: how are studies being designed and reported? An update on current practice in the published literature between 2005 and 2008. Patient 2010; 3: 24956.
  • 17
    Fraenkel L. Conjoint analysis at the individual patient level: issues to consider as we move from a research to a clinical tool. Patient 2008; 1: 2513.
  • 18
    Fraenkel L. Feasibility of using modified adaptive conjoint analysis importance questions. Patient 2010; 3: 20915.
  • 19
    Fraenkel L, Bogardus ST, Concato J, Felson DT, Wittink DR. Patient preferences for treatment of rheumatoid arthritis. Ann Rheum Dis 2004; 63: 13728.
  • 20
    Llewellyn-Thomas HA. Values clarification. In: Edwards A, Elwyn G, editors. Shared decision-making in healthcare: achieving evidence-based patient choice. 2nd ed. Oxford: Oxford University Press; 2009.
  • 21
    Fraenkel L, Rabidou N, Wittink D, Fried T. Improving informed decision-making for patients with knee pain. J Rheumatol 2007; 34: 18948.
  • 22
    Kopec JA, Richardson CG, Llewellyn-Thomas H, Klinkhoff A, Carswell A, Chalmers A. Probabilistic threshold technique showed that patients' preferences for specific trade-offs between pain relief and each side effect of treatment in osteoarthritis varied. J Clin Epidemiol 2007; 60: 92938.
  • 23
    Peters E, Hart PS, Fraenkel L. Informing patients: the influence of numeracy, framing, and format of side effect information on risk perceptions. Med Decis Making 2011; 31: 4326.
  • 24
    National Institutes of Health. Handout on health: osteoarthritis. National Institute of Arthritis and Musculoskeletal and Skin Diseases; 2006.
  • 25
    Treatment choices for knee osteoarthritis: decision aid summary. 2010. URL: http://decisionaid.ohri.ca/AZsumm.php?ID?1087.
  • 26
    User manual: decisional conflict scale (16 item statement format). Ottawa: Ottawa Hospital Research Institute; 1993, updated 2010. URL: http://decisionaid.ohri.ca/docs/develop/User_Manuals/UM_Decisional_Conflict.pdf.
  • 27
    Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987; 40: 37383.
  • 28
    Roos EM, Toksvig-Larsen S. Knee Injury and Osteoarthritis Outcome Score (KOOS): validation and comparison to the WOMAC in total knee replacement. Health Qual Life Outcomes 2003; 1: 17.
  • 29
    Zikmund-Fisher BJ, Couper MP, Singer E, Ubel PA, Ziniel S, Fowler FJ Jr, et al. Deficits and variations in patients' experience with making 9 common medical decisions: the DECISIONS survey. Med Decis Making 2010; 30: 85S95S.
  • 30
    Elwyn G, O'Connor A, Stacey D, Volk R, Edwards A, Coulter A, et al. Developing a quality criteria framework for patient decision aids: online international Delphi consensus process. BMJ 2006; 333: 417.
  • 31
    Elwyn G, O'Connor AM, Bennett C, Newcombe RG, Politi M, Durand MA, et al. Assessing the quality of decision support technologies using the International Patient Decision Aid Standards instrument (IPDASi). PLoS One 2009; 4: e4705.
  • 32
    Spunt BS, Deyo RA, Taylor VM, Leek KM, Goldberg HI, Mulley AG. An interactive videodisc program for low back pain patients. Health Educ Res 1996; 11: 53541.
  • 33
    Weng HH, Kaplan RM, Boscardin WJ, Maclean CH, Lee IY, Chen W, et al. Development of a decision aid to address racial disparities in utilization of knee replacement surgery. Arthritis Rheum 2007; 57: 56875.
  • 34
    Ringash J. Review: decision aids increase patients' knowledge and realistic expectations and reduce decisional conflict. Evid Based Med 2004; 9: 21.
  • 35
    Lloyd FJ, Reyna VF. Clinical gist and medical education: connecting the dots. JAMA 2009; 302: 13323.
  • 36
    Losina E, Walensky RP, Kessler CL, Emrani PS, Reichmann WM, Wright EA, et al. Cost-effectiveness of total knee arthroplasty in the United States: patient risk and hospital volume. Arch Intern Med 2009; 169: 111322.
  • 37
    Arden N, Nevitt MC. Osteoarthritis: epidemiology. Best Pract Res Clin Rheumatol 2006; 20: 325.