Abstract
Background
Antipsychotics are widely used to treat symptoms of dementia. A nationwide study revealed that 22% of residents in nursing homes were prescribed antipsychotics. 1 The Centers for Medicare & Medicaid Services (CMS) found that in 2010, almost 40% of nursing home residents with signs of dementia received at least 1 antipsychotic. 2 However, the increased risk of serious adverse outcomes associated with antipsychotic use prompted the Food and Drug Administration (FDA) to issue a black box warning highlighting increased mortality risk when they are used for dementia-related psychosis. 3 Despite the evidence of serious risks and the Omnibus Budget Reconciliation Act of 1987 regulations (OBRA-87) targeting inappropriate psychotropic drug use in nursing homes, a chart review by the Department of Health and Human Services Office of Inspector General estimated that 22% of antipsychotics prescribed to nursing home residents were used without proper indication, monitoring, or dosing.4,5 The American Medical Directors Association (AMDA) and the American Geriatric Society (AGS) highlighted antipsychotic use in dementia in their Choosing Wisely campaigns. 6 They recommended that antipsychotics should not be used as first-line treatments for behavioral and psychological symptoms and should not be prescribed without an assessment for an underlying cause of behavior. 6 In response to the increased use of antipsychotics in people with dementia, the CMS implemented an initiative in 2012 to reduce antipsychotic use in nursing homes by 15%, a target that was met in 2014. 7 The goal was updated to a 25% reduction by the end of 2015 and a 30% reduction by the end of 2016. 8
Educational interventions that target inappropriate prescribing practices of antipsychotics in nursing homes have shown some success in reducing antipsychotic use, particularly when they have targeted nursing home staff in addition to prescribers.9,10 Recent efforts by CMS and others to reduce antipsychotic use have focused largely on the use of nondrug approaches, including person-centered care instead of antipsychotics to manage behavioral symptoms. 7 Another focus has been on identifying potential environmental, physical, or psychological causes leading to behavioral symptoms, such that interventions can directly target these causes. Implementation of these best practices may prevent unnecessary antipsychotic use and potentially improve the well-being and safety of individuals with dementia, but education is necessary to support physicians who are less familiar with these approaches to care. Understanding physician preferences and needs is the first step in developing targeted educational resources for this group that aim to have a meaningful impact on practice.
Previous work has explored physician preferences for information. A meta-analysis from 1997 found that books, journals, meetings, drug companies, workshops and courses were among the most preferred sources of information. 11 Other work showed that primary care providers in another rural state used print resources more often than online resources, though both were commonly used. Colleagues were also common sources of information. 12 This is consistent with a review from 2006 concluding that colleagues and paper-based resources were the primary information sources for physicians. 13 More recent studies have highlighted increasing use of web-based sources,14,15 which may relate to the ubiquity and easy access to computers and portable digital devices in the current practice environment. Source preferences may also vary by digital literacy and trust in the information provided, and specific information needs. 16
The purpose of this study was to assess the needs and preferred resources of Iowa family physicians to inform the development of educational resources aimed at supporting appropriate use of antipsychotics in people with dementia. A secondary goal was to analyze the differences and similarities in the responses of family physicians that were nursing home medical directors to those who were not medical directors (nonmedical directors). We anticipated that medical directors would report greater use of and confidence in nondrug management strategies and greater satisfaction with training and knowledge regarding dementia care, given their higher level of involvement in the provision of nursing home care. We also examined whether medical directors reported different patterns of antipsychotic prescribing in their patients or different levels of awareness of safety issues with antipsychotics, anticipating that they may have greater awareness of risks and choose evidence-based treatments more often. Finally, we explored whether medical directors and nonmedical directors had different preferences for information sources. We considered the possibility that medical director experience or greater interactions with nursing home staff could influence their preferences for information sources, though we had no particular expectations about what differences might exist.
Methods
This study was reviewed and approved by the Institutional Review Board at The University of Iowa. A list of names and addresses of 1256 family practice physicians licensed in Iowa was obtained from the Iowa Board of Medical Examiners. A cover letter, prepaid return envelope, and questionnaire were mailed to 500 randomly selected physicians. One follow-up questionnaire was mailed to nonresponders 3 weeks after the initial mailing.
Instrument
A 17-item questionnaire (see the appendix) titled, “Needs Assessment: Caring for People with Dementia and Challenging Behaviors” was developed by the research team to assess physician needs in caring for patients with dementia and dementia-associated behavioral symptoms. The survey included items on the following: the current techniques and resources used to care for patients with dementia, knowledge and satisfaction with dementia training, which potential resources and types of information would be helpful when caring for patients with dementia, and useful means to communicate with family members about approaches to care. Questions relating to satisfaction with training, awareness of the black box warning for antipsychotics, and confidence in managing patients with non-drug approaches were answered using a 7-point Likert-like scale. Demographic information, including age, gender, ethnicity, race, as well as medical director status was also queried. The questionnaire was not piloted due to overall project timelines but was reviewed by a large research team, including a number of physicians.
Statistical Analysis
Two research team members entered questionnaire responses into separate Excel files, and a verification program was run to check for errors. Data were analyzed in SAS 9.3 (SAS Institute, Cary NC). For demographic and qualitative survey questions, the number and percentage of physicians selecting each answer were calculated. Responses were stratified by medical director status. Differences in the 2 groups’ responses were evaluated by performing chi-squared tests, Fisher exact tests, and Wilcoxon rank-sum tests as appropriate.
Results
A total of 101 (20%) completed questionnaires were returned and 2 were undeliverable of the 500 mailed questionnaires. The mean age of responding physicians was 52 years ranging from 33 to 75 years. The majority were white (98%), non-Hispanic (94%), and male (76%). All the physicians reported they cared for a mean of 25 (SD = 30.5) nursing home patients ranging from 0 to 200 patients in an average of 3 nursing homes (see Table 1). These physicians had a mean of 21 years in practice ranging 3 to 50 years.
Medical Directors and Nonmedical Directors Respondent Characteristics and Nursing Home Patients Cared for.
Percentages based on those who responded to this question (38 missing).
P < .05.
Forty (40%) of the physicians were nursing home medical directors. They were significantly older than physicians who did not serve as medical directors. The medical directors cared for significantly more patients with dementia and cared for more nursing home patients at a larger number of nursing homes compared with the nonmedical directors (see Table 1).
Both medical directors and nonmedical directors reported moderate satisfaction with training in managing challenging behaviors, felt nursing staff and family influence on use of specific medications was moderate, and had moderate confidence in use of nondrug management strategies, with no statistically significant differences between groups. Medical directors were significantly more aware of the black box warning for antipsychotics as compared with nonmedical directors (P = .03). Substantial variability in the use of nondrug approaches instead of medications was observed, with medical directors more likely to be using nondrug approaches (see Table 1).
Physicians obtained information about challenging behaviors in patients with dementia from a variety of sources. UpToDate and consulting pharmacists were the most frequent resources used (see Table 2). Medical directors were less likely to use the online resource UpToDate to get information on challenging behaviors associated with dementia. Forty-four percent of the physicians used a consulting pharmacist. Eighty-two percent of the participants wanted to learn more about managing difficult behaviors without drugs, 68% about caring for difficult patients, 67% about medications, and 59% about tools for family members. Other topics and kinds of information the participants wanted to learn about are listed in Table 2. The preferred sources of new information or clinical tools to help them and their staff deal with challenging behaviors associated with dementia included online resources (54%), pocket guides (37%), handbooks (34%), consulting pharmacists (32%), and facility in-services (31%). Additional preferred sources of new information or clinical tools are listed in Table 2.
Sources Used and Additional Information Wanting to Learn on Challenging Behaviors Associated With Dementia.
Abbreviation: PDA, personal digital assistant.
Other sources of information listed were Epocrates, American Academy of Family Physicians, Dynamed, Medline, Medscape, and Google.
Table restricted to responses endorsed by greater than 10 physicians. Other options endorsed by at least one physician included informational poster and other. Screensavers and Facebook or other social media were not endorsed by anyone.
P = .01.
To facilitate communication with the families of people with dementia regarding medications and nondrug approaches, participants primarily preferred 1-page handouts (76%), followed by brochures (39%), easier-to-read materials (38%), and booklets (34%) (Table 3). No significant differences were noted between medical directors and nonmedical directors regarding communication with families.
Preferred Resource/Source of Information to Help Communicate With Families About Medications and Nondrug Approaches. a
There were no statistically significant differences between groups.
The antipsychotics most often used in practice by respondents were risperidone (88%), quetiapine (86%), and haloperidol (78%) (Table 4). Significantly more medical directors used haloperidol compared with nonmedical directors, while significantly more nonmedical directors used ziprasidone compared to medical directors.
Antipsychotics Used in Practice by Surveyed Physicians. a
Table includes antipsychotics used by at least 10 respondents. Other antipsychotics used in at least one respondent’s patients included paliperidone, thioridazine, fluphenazine, thiothixene, chlorpromazine, trifluoperazine, asenapine, perphenazine.
P < .05.
Discussion
The results of this study highlight the educational needs and resource preferences of Iowa family physicians regarding the care of people with dementia and challenging behavioral symptoms. There were no statistically significant differences in the types of resources used or requested by medical directors and nonmedical directors, with the exception of more use of UpToDate by nonmedical directors. UpToDate is a popular resource for getting current information about diagnosis and management of health conditions. Nonmedical directors may have used it more often for this purpose because of less familiarity with management of behavioral symptoms. However, this mirrored preferences for online resources, which could also relate to differences in age distributions and experiences using these sources in training.
Our findings highlight that participating nursing home medical directors were significantly older, worked with more patients who have dementia, and were more aware of the FDA warnings on antipsychotic use in patients with dementia than physicians who were not medical directors. However, they were not necessarily more confident in their abilities to treat and manage challenging behaviors associated with dementia without drugs and were not more likely to use nondrug approaches instead of medication to manage behavioral symptoms of dementia than physicians who were not medical directors. Levels of confidence in non-drug management strategies were moderate among both medical directors and nonmedical directors. Previous work has consistently identified management of behavioral and psychological symptoms in people with dementia as a challenge for primary care providers.17-22 Only about a third of providers expressed confidence in managing these symptoms in one study from the United Kingdom. 17 Over 50% of nonmedical directors reported using nondrug approaches instead of medications in half or fewer of their patients with dementia and behavioral symptoms, and 45% of medical directors responded the same. The survey did not explore the process by which physicians made decisions to use medications to manage symptoms, such as the use of nondrug interventions prior to or concurrent with prescribing medications, or the duration of time for which medications were prescribed. This information would have been useful to provide more context about medication use. As medical directors reported they have more nursing home patients and patients with dementia and provide assistance at more nursing homes than nonmedical directors, a higher level of confidence among medical directors was expected. These moderate confidence levels may reflect the perception that nondrug management of behavioral symptoms is a nursing task, and nursing home staff contacts physicians when non-drug approaches have failed. They may also reflect the severity of symptoms in the patients for whom they care, as neuropsychiatric disturbances are frequent causes of nursing home placement. 23 However, other responses indicate these physicians were interested in learning more about nondrug management of behavioral symptoms, suggesting a need for education for both medical director and nonmedical director family physicians.
A majority of surveyed physicians also wanted more information on medications used for the treatment of dementia. Our survey showed medical directors were more likely to use haloperidol and nonmedical directors were more likely to use ziprasidone. This may suggest that medical directors are more informed about evidence on antipsychotic efficacy in dementia. Clinical trials have shown a reduction in patient aggression with the use of haloperidol in dementia, whereas no randomized controlled trials have evaluated the safety and efficacy of ziprasidone in patients with dementia. It may also reflect the medical director’s older age, as they may be more likely to have used haloperidol, an older antipsychotic, in their training. 24 Alternatively, it could reflect a more cost-conscious approach to prescribing, given that haloperidol is a relatively low-cost generic drug. With regard to ziprasidone, nonmedical directors were younger and may have had more experience with ziprasidone in their training or been influenced by sales representatives, as it is a newer medication. It is also possible that they find some value in its use. It is not clear that it does not work for this indication even though evidence from randomized controlled trials is lacking. Regardless, we believe our findings suggest that continuing education on antipsychotic selection was needed even among medical directors, as evidenced by the common use of quetiapine. In 4 randomized controlled trials, quetiapine showed no greater efficacy than placebo in treating neuropsychiatric symptoms of dementia. 25 One trial did demonstrate evidence of benefit in a secondary analysis at a dose 200 mg/d, which is higher than the CMS allowable dose for chronic use in nursing home residents with dementia.26,27 Various sources suggest quetiapine is the most commonly used antipsychotic in US nursing homes.1-4 It is difficult to resolve its frequent use given that so much evidence suggests it is not effective for management of neuropsychiatric symptoms of dementia. 25
The information provided by this survey on the sources of information preferred by physicians is somewhat unique, in part because technology has changed the way that healthcare providers can access information. While this survey did not directly evaluate preferences for in-person meetings, workshops, and conferences, historically among the commonly preferred sources of information, 11 the results show that online resources such as UpToDate are commonly used to obtain information, and that online resources were the most preferred source of new information among the options provided. This is consistent with research highlighting a shift toward electronic resource utilization.14,15 About a quarter of respondents included mobile device applications among their preferred sources. Many physicians also endorsed paper-based tools such as pocket guides and handbooks, however, as well as obtaining information from colleagues such as consultant pharmacists. The frequent endorsement of pocket guides as a preferred source of information may reflect the condensed but useful information such sources can provide. Prior work on printed education materials found that primary care providers preferred simple concise documents with links to sources that provide detailed information. This survey did not directly evaluate whether physician colleagues were a preferred source of information. It is interesting that consulting nurses were not commonly used sources of information or preferred as sources of new information, given that consulting nurses practicing in the nursing home setting might be expected to be knowledgeable about managing dementia symptoms without drugs. The results could simply reflect that responding physicians had limited experience working directly with consulting nurses. This could relate to infrequent use of consulting nurses. Another possibility is that consulting nurses interact more often with nursing home staff delivering behavioral interventions than with medical staff. Nursing home staff may involve physicians primarily when they perceive a need for a medical assessment or medication and involve them less in decisions about nondrug interventions. About a quarter of respondents reported receiving information on managing challenging behaviors from drug company representatives, which on the surface seems less than ideal but could reflect discussions of risks. Finally, physicians also indicated that 1-page handouts for were their most preferred source of information to provide to patients.
Limitations of this study include a relatively small sample size of 101 physicians with only a 20% response rate. Results may not be representative for family practice physicians in Iowa as a whole, or generalizable to other states. Those who chose to complete the survey may have been more interested in the topic. Respondents may have had difficulty interpreting some questions.
Conclusions
The responses to this survey were used to help inform the development of the Improving Antipsychotic Appropriateness in Dementia Patients (IA-ADAPT) educational program and resources. 28 This program provides information through online lectures, pocket guides, a web-based mobile device application, evidence reviews, and other resources to assist health care providers involved in management of behavioral and psychological symptoms of dementia. It also includes a reader-friendly shared decision-making guide for patients and families who may be involved in decisions about antipsychotic use and symptom management. Participants did not strongly endorse all these sources in the formats provided, but the pocket guides supported by more detailed information provided on the website are consistent with preferences for printed materials, and the program as a whole is consistent with preferences for online information. These resources are available at no cost at https://www.healthcare.uiowa.edu/igec/iaadapt/. Although this survey was specific to the care of patients with dementia, the preferences for sources and types of information may also be informative when developing educational resources and interventions to promote best practices in caring for other conditions.
Footnotes
Appendix
Acknowledgements
All statements in this report, including its findings and conclusions, are solely those of the authors and do not necessarily represent the views of the Patient-Centered Outcomes Research Institute (PCORI), its Board of Governors or Methodology Committee, or the Agency for Healthcare Quality and Research. The authors would like to thank the other members of the “IA-ADAPT: Improving Antipsychotic Appropriateness in Dementia Patients” team for their input on survey design, and the study participants for taking the time to complete and return the survey.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by R18HS019355 from the Agency for Healthcare Research and Quality, with additional support provided by a Patient Centered Outcomes Research Institute (PCORI) Program Award (1131).
