Date: March 10th, 2016

Guest Skeptic: Dr. Tim Platts-Mills. Tim is an Assistant Professor of Emergency Medicine; Director, Clinical Research; Co-Director of Geriatric Emergency Medicine at the University of North Carolina. His career goal is to improve the quality of emergency care for older adults through research, research mentorship, and support of the larger community of geriatric emergency medicine researchers. Tim’s research group is currently developing a protocol to screen for elder abuse in the emergency department. He also serves as a decision editor in geriatrics for Annals of Emergency Medicine.

Case: 78-year-old female with history of osteoporosis trips and falls while rushing between flights at the airport. She arrives to your emergency department with right wrist pain and after your evaluation you order an x-ray that reveals a Colles’ fracture. No other injuries are identified. After you assess her response to emergency department analgesia (acetaminophen) and splint the fracture, you prepare to discharge her so that she can catch the next flight home and you contemplate the pros and cons of various pain medications. Personally, if you suffered a wrist fracture, you would favor whatever analgesic provided sufficient relief to return to daily activity, and nothing more than that. You decide to ask the patient for her pain management preferences, but you wonder what the role is for emergency department shared decision-making in geriatric patients.

Background: Most clinicians are familiar with informed consent or informed decision making but may not be as familiar with the concept of shared decision making (SDM). Valerie Billingham in 1998 is credited with proclaiming “nothing about me without me” at the Salzburg Global Seminar. This statement succinctly captures the vision that medicine must always consciously respect human dignity. Her statement is credited as the genesis for “Shared Decision Making” (SDM) in medical decision-making.

  • Key elements to SDM (Barry and Edgman-Levitan NEJM 2012):
    • Patient and the doctor collaborate on reaching a decision about a management strategy for a given problem
    • It first requires a situation where more than one reasonable option exists
    • SDM also requires that a patient be given the information they need to choose among the competing acceptable strategies
    • It shifts focus from “disease” and towards understanding patients’ experience of illness
    • Barry states that SDM is the pinnacle of patient-centered care

The Affordable Care Act of 2010 in the United States emphasizes SDM that includes communication strategies to help patients collaboratively choose the best treatment option.

Although the 3-prongs of Evidence Based Medicine include research evidence, clinician expertise, and patient preferences, medicine has too often underemphasized the unique perspectives of patients and caregivers.

Clinical Question: Does shared decision-making for analgesic selection in older adults discharged home from the emergency department with acute musculoskeletal pain improve pain relief?

Reference: Holland et al. A Prospective Evaluation of Shared Decision-Making Regarding Analgesics Selection for Older Emergency Department Patients with Acute Musculoskeletal Pain. Acad Emerg Med 2016

  • Population: Adults >60 years old with acute, moderate-to-severe (pain greater than or equal to 4/10 on a 0-10 scale) musculoskeletal pain discharged home from the emergency department.
    • Exclusion: Cognitive impairment (Six Item Screener score of 3 or less), pain for >1 month, daily use of opioid pain medication prior to current pain onset, headache, chest pain, or abdominal pain, lack of telephone for follow-up, or non-English speaking.
  • Intervention: No intervention, but rather a descriptive, prospective, convenience-sample observational study of patient’s preferences for shared decision-making, perceptions of shared decision-making with their musculoskeletal pain related ED visit, amount of analgesic information received, and ED care satisfaction within 24 hours of discharge via telephone. Another telephone interview at 1-week assessed effectiveness of pain relief and functional recovery.
  • Comparison: No comparator group.
  • Outcome:
    • Primary Outcome: Change in pain severity from the time of ED arrival to the 1-week follow-up phone interview.
    • Secondary Outcomes: Satisfaction with the decision made in the ED about how to treat pain at home and satisfaction with the recommended or prescribed pain medication, and medication side effects.

Author’s Conclusions: “In this sample of older adults with acute musculoskeletal pain, the reported desire of patients to contribute to decisions regarding analgesics varied based on both patient and provider characteristics. SDM was not significantly related to pain reduction in the first week or type of pain medication received, but was associated with greater patient satisfaction.”

Quality Checklist for Observational Trials:checklist

  1. Did the study address a clearly focused issue? Yes. Are patients’ perceptions of shared decision-making associated with positive or negative outcomes in acute pain management for musculoskeletal pain in older adults?
  2. Did the authors use an appropriate method to answer their question? Unsure. Although a randomized controlled trial that measured the fidelity and methods used to provide SDM at the bedside would yield more definitive cause-effect relationships, this observational study does give preliminary evidence that can serve as hypothesis-generating data for future studies.
  3. Was the cohort recruited in an acceptable way? No. Convenience sampling so uncertain selection bias, as noted by the authors.
  4. Was the exposure accurately measured to minimize bias? No. There was no assessment of if or how shared decision-making actually occurred in the ED. The study only evaluates patients’ perceptions of shared decision-making.
  5. Was the outcome accurately measured to minimize bias? Yes. The authors used validated instruments (although not validated in ED settings or amongst ED patients) to assess individual patient’s preferences for sharing decisions with healthcare providers.
  6. Have the authors identified all important confounding factors? Unsure. The authors evaluate for the most important confounders in shared decision-making for older adults (cognitive impairment) and for all patients (health literacy), but as noted below they use screening instruments that appear to less accurately identify dementia and limited health literacy.
  7. Was the follow up of subjects complete enough? No. Only 94/157 (60%) had one-week follow-up and were included in the analysis with those lost to follow-up more likely to be female and black.
  8. How precise are the results/is the estimate of risk? Unsure. Precision estimates for the primary outcome are not provided.
  9. Do you believe the results? Yes. The results indicate variable interest in older adults to engage in shared decision-making and the demographics of this population (including limited health literacy in 27%) match those of other ED studies.
  10. Can the results be applied to the local population? Yes. The variable acceptance of shared decision-making in outpatient pain management decisions probably reflect the general geriatric ED population of most sites, but the implications of this research (how to use it to improve patient-centric outcomes) is unclear.
  11. Do the results of this study fit with other available evidence? Unsure. Little research exploring ED shared decision-making in older adults exists, but the demographics of this population match that of other geriatric population studies in the ED.

Key Results: Patients were mostly female (62%) with mean age 70 years and 74% were white and in severe pain (69%) at triage. There was an overall mean pain score reduction of 2.1 (6.6 to 4.5) between the ED visit and 1-week follow-up.

  • Preferences for SDM were:
    • Active (16%): dering patient’s opinions or wanted to leave all treatment decisions to the doctor. Patient makes the decision independently or make the decision after seriously considering input from the doctor.
    • Collaborative (37%): Share the decision with the doctor.
    • Passive (47%): Have the doctor make the final decision about treatment after considering patient’s opinions or wanted to leave all treatment decisions to the doctor.

Characteristics associated with greater desire for active role in decision making included: college graduate, care received from nurse practitioner, and care received from a female provider.

No significant association between 1-week pain improvement and any perceived degree of SDM was noted.

In addition, no difference in number of analgesic medication side effects or type of pain medication (acetaminophen, NSAID, or opioid) was observed.

However, patients who perceived receiving more SDM noted more satisfaction with the pain medication that they received (2.7 vs. 3.9 on a 1-5 Likert scale with 5 representing “completely satisfied”).

Screen Shot 2015-04-25 at 3.11.12 PM

Dr. Chris Carpenter

Dr. Chris Carpenter

Adults over age 65 represent 18%-20% of ED patients in most hospitals nationwide. Surveys of ACEP membership in 1992 and in 2006 indicated that EM providers believe geriatric patients are more challenging to diagnose and manage, and consume more time and resources. Yet older adults often leave the ED feeling dissatisfied with the care received. Tim and I co-authored GED Guidelines that have been endorsed by ACEP, SAEM, AGS, ENA, and last week by CAEP in an effort to attain the Triple Aim for these patients (improved healthcare experience at same or lesser cost with improved outcomes). Prompt, effective analgesia is obviously essential for an improved patient experience.

  1. Screening Tool for Occult Cognitive Dysfunction: One of the confounders for geriatric SDM is the presence of occult cognitive dysfunction. Multiple studies indicate that delirium and dementia are usually unrecognized in ED elderly.   Therefore, these authors used the Six Item Screener as one exclusion criterion in order to reduce the impact of occult cognitive dysfunction. However, multiple ED validation studies have demonstrated that the Six Item Screener is inaccurate to either rule-in or rule-out occult dementia (positive likelihood ratio [LR+] =3.3, negative likelihood ratio [LR-]= 0.33). More accurate dementia screening instruments like the AD8, the Ottawa 3DY, or the Short Blessed Test, could be used in the future (Carpenter et al and Wilding et al). In addition, ultra brief screening instruments for acute delirium could also be used (Han et al and  Han et al).
    • RESPONDS: Ottawa 3DY asks patients to report the day of the week, the year, and the date, plus spelling the word WORLD backward, with a three or less being cognitively impaired. I usually check my phone to see what the date is, sometimes multiple times a day. So…I’d probably fail this. Six item screener uses day of the week, the year, and the month (plus remembering apple, table, and penny), which seems a bit more friendly, plus you are allowed to miss two. For a clinical trial, the question is not how really well you can measure it but whether we think SDM could help patients with mild cognitive impairment? I think the answer is yes, so I’d pick a measure which is more inclusive for deciding who gets included in the study. Delirium is definitely missed in the ED, but I think less so in patients with acute MSK pain. Probably would be good to screen for delirium and exclude in the trial.
  2. Health Literacy: Another patient-level factor that can negatively impact the effectiveness of SDM is health literacy, which is based upon the gold standard of the S-TOFHLA is limited in about 25% of urban ED patients in the United States. Studies have shown that ED clinicians do not detect health literacy subjectively. Furthermore, health literacy is exacerbated by cognitive dysfunction so in older adult studies, both health literacy and cognitive impairment should be measured concurrently as these authors did. However, the choice of the REALM-R (LR+ = 2.1, LR- = 0.3) is not the best choice based on ED research indicating that the Newest Vital Sign is more accurate to identify a subset of patients at less risk of limited health literacy with LR+ 1.8 and LR- 0.04.
    • RESPONDS: The Newest Vital Sign was developed by a team including physicians at UNC (Dewalt, Pignone). We initially considered using the Newest Vital Sign (NVS) as an assessment of health literacy, but noticed that it was much more difficult and time-consuming to administer in the ED compared to the REALM-R. The NVS questionnaire includes six questions based on the information provided in an example nutrition label. Some of the questions are rather long and require basic mental math and reasoning skills, which may be too intensive for regular use in the ED. Additionally, the NVS is reported to take over three minutes on average to complete for adults aged 18 and older. This is much longer than the REALM-R, which can be easily finished in less than 30 seconds since it only involves reading ten words aloud. There is no free lunch – if you use a more accurate tool that takes 3 minutes and has 6 word problems, you are likely to lose some patients, which introduces selection bias. I think REALM-R was good choice in this case.
  3. Descriptive Studies: Descriptive studies are limited by the measures employed to capture, define, and characterize a phenomenon. The current study tries to evaluate a nebulous activity called “shared decision making”, a process that many experts still cannot reach consensus about how to define. They use the Control Preferences Scale to characterize the extent to which patients wish to lead or follow in reaching a medical decision with physicians. The authors then use the Shared Decision-Making Questionnaire (SDM-Q-9) to explore patient’s perceptions of how effectively or ineffectively their ED provider used SDM while managing their current musculoskeletal pain. As with any instruments (see the studies about dementia, delirium, and health literacy referenced above), the performance of measures that work in primary care clinics, post-op settings, or nursing homes may differ markedly from how unwell they perform in ED settings. Using appropriate methods, instruments like these should be validated in ED settings before they can be confidently applied and interpreted in ED settings. For example, the SDM-Q-9 assesses only perceptions of SDM, but does not try to evaluate what actually occurred. The authors note that the “Observing Patient Involvement in Decision Making” (OPTION) scale assesses SDM using a third person or video camera thereby removing the uncertainties of patient memory or subjective interpretations. SDM is inherently subjective, though, so even if pristine methods were used by providers to engage willing and able patients in SDM, if the patient’s next day perceptions are that SDM attempts were sub-optimal than it may not matter what a third party observer believes.
    • RESPONDS: It would be interesting to watch what happens in patient rooms, and try to do objectives assessments of SDM. However, to quote Eisenhower, I would make the problem bigger here. The tool we used to measure SDM is a problem, but the presence of unmeasured or poorly measured confounders is also a problem. In some lines of research, you only have observational studies. You can’t randomize patients to smoking or not smoking to see an effect on cancer. For understanding whether an SDM approach improves outpatient pain management for older adults, we can do clinical trials, which would allow us to overcome many of these limitations. So…I am happy to accept all the limitations described, including this particular one about the SDM-Q-9. The real value of this research, is I hope, that it inspires a research group (perhaps ours, perhaps another group) to do a clinical trial to evaluate SDM for the outpatient treatment of acute musculoskeletal pain in older adults. I think it also provides some of the background needed to design such a trial in terms of which patients might be included and what outcomes might be considered. I will add however, that SDM in this setting has a ton of face validity. We know there are pain medication options, we know these medications have risks, and we know that physicians often don’t get enough information from patients about what they should or shouldn’t take and many patients don’t even know the different between acetaminophen and ibuprofen.
  4. Adjusting for Multiple Covariates: Another aspect of this study that leave significant uncertainty are the attempt to adjust for multiple covariates (age, gender, race, initial pain severity, health literacy) with a sample size of 94.
    • RESPONDS: Agreed. As you start to adjust for covariates you need a bigger sample. In the article we showed unadjusted and adjusted estimates – and they are fairly similar – which makes me think that there wasn’t a lot of confounding going on. The biggest confounder is probably education – educated patients are more likely to engage in SDM and also probably more likely to do a number of behavioral things to make them recover (like get good sleep, stay active, take medications appropriately).
    • In addition, the authors made no attempt to educate providers about techniques or barriers to SDM, nor how to evaluate patient comprehension. For example, how did individual providers communicate comparative effectiveness estimates for acetaminophen vs. NSAIDS vs. opioids? Did they use studies or generalized gestalt?
    • RESPONDS: Right – We didn’t intervene on providers. If we could effectively teach SDM to all our ED providers, that would probably be a good thing for patients, but we wouldn’t have any variance in the exposure so we wouldn’t know if it was helpful. In truth, we have an outstanding group of physicians at UNC – I think as good as anywhere in the country. But, most providers are not giving much information to patients about these options.
  5. Statistical vs. Clinical Significance: Finally, the adjusted difference in “satisfaction with pain medication” (2.7 in low SDM vs. 3.9 in high SDM, p=0.006) may be statistically significant, but whether this is clinically relevant or important to patients is unknown.
    • RESPONDS: Also a good question. Having a satisfaction score of 3 meant they were “somewhat satisfied”; having a score of 4 meant the patient was “quite a bit satisfied.” Seems like an important difference for this outcome, but not sure. I agree with the broader sentiment implied here, that for the clinical trial, I’d like to see an impact on pain and function, which strike me as much more important than satisfaction with the medication.

Comment on authors conclusion compared to SGEM Conclusion: We agree with the authors’ conservative conclusions. In this single center, convenience sampling of older adults without overt cognitive impairment and with acute or sub-acute musculoskeletal pain, the majority of patients desire some SDM in selecting outpatient analgesia. Perceived receipt of ED SDM is associated with improved patient satisfaction about the analgesic prescribed and is not associated with an increased use of opioids, but is not associated with either less pain or less medication side effects at 1-week.

SGEM Bottom Line: SDM in selecting pain relief medications in older adult musculoskeletal pain patients is preferred to varying degrees by most patients, but in this study is not associated with faster recovery (less pain), less side effects, or a predisposition to any particular analgesic.

Case Resolution: You spend 5-minutes discussing the side effects and mechanism of action of acetaminophen, NSAIDS, and opioids with this patient. Since her pain is minimal with her fracture splinted and because she already has problems with intermittent constipation, she is not interested in opioids (even with a bowel regimen). She collaboratively decides to use acetaminophen (which was effective in the ED) and you discharge a comfortable, contented patient to catch the next flight home.

Clinically Application: Uncertain. Since the authors did not assess the actual delivery of SDM real-time in the ED or provide any control group, the cause-effect relationship of SDM for these patient-centric pain outcomes cannot be elucidated by these results. How and when SDM should be used on whom and by which ED personnel cannot be determined by this study design. As the authors note, a clinical trial is required to confirm this benefit (as well as to better understand the how/when/who issues of SDM delivery). The objectives and priorities of these future studies will be the focus of the 2016 Academic Emergency Medicine Consensus Conference on May 10th, 2016 in New Orleans that every SGEM listener should plan to attend!

What do I tell my patient? Shared Decision-Making is the process by which patients and healthcare providers mutually review treatment options in deciding upon the optimal choice for the individual patient. This small study indicates that in older adults with acute pain in the ED, patients who receive SDM in selecting a pain medication are more satisfied with the choice of pain medicine received, but do not obtain faster pain relief or less side effects. Would you like me to review the effectiveness and side effects of different pain medications available?

Keener Kontest: Last weeks’ winner was Chip Lange. Chip is a PA from Missouri.  He the knew the Lego Movie had cyanoacrylate glue know as the Kragle as a superweapon?

Listen to the podcast for this weeks’ question. If you think you know the answer sent an email to TheSGEM@gmail.com with “keener” in the subject line.  The first correct answer will receive a cool skeptical prize.

Now it is time for the SGEMers to join the conversation. What do you think about this #SGEMHOP episode? What questions do you have for Tim and his team? Reach out to us on twitter, Facebook or the SGEM blog. The best social media feedback will be published in AEM.

Remember to be skeptical of anything you learn, even if you heard it on the Skeptics’ Guide to Emergency Medicine.

 Screen Shot 2016-01-10 at 11.16.25 AM