ORIGINAL STUDIES Profile of Physicians in the Nursing Home: Time Perception and Barriers to Optimal Medical Practice Thomas V. Caprio, MD, Jurgis Karuza, PhD, and Paul R. Katz, MD Objectives: To describe physician medical practice in nursing homes, including actual time spent for routine encounters with nursing home residents and demographic characteristics of the physicians who serve as medical directors; to determine the congruence between actual time spent for routine encounters with nursing home residents and the physician s view of the optimal time; and to identify barriers to providing optimal visits. Design: A mail survey of a national random sample of 200 medical directors of all Medicare-certified nursing facilities using the Dillman Total Design mail survey methodology. Participants: 100 medical directors (50% response rate). Measurements: The survey consisted of open- and closed-ended items on the following: the demographic characteristics of the medical director; demographic characteristics of the nursing home; the extent of the medical director s nursing home practice, including the ideal and actual time spent in nursing home visits for 4 common types of visits; and perception of barriers to providing optimum visits in the nursing home. Results: Medical directors were most likely to be primary care physicians, the majority of whom were male; had practiced in long-term care for more than 18 years; were medical directors in 2 facilities; provided, on average, primary care in 4 facilities; spent 31 hours per month in the nursing home with nursing staff; and devoted 44% of their practice to nursing homes. Most, (74%) were members of the American Medical Directors Association (AMDA), 41% were certified medical directors (CMD), 42% had a certificate of added qualification (CAQ) in geriatrics, and only 15% had fellowship training. Reports of actual time spent on 4 common types of nursing home visits were significantly less than optimal visit times, but fellowship-trained physicians reported significantly greater discrepancies between the optimal and actual time spent for the 30- to 60-day reviews and readmissions compared with physicians who were not. A parallel pattern was seen comparing physicians with and without CAQs. Nursing support and accurate/accessible information were recorded as most problematic and reimbursement the least problematic barrier to providing optimal nursing home visits. Conclusion: The present study provides a snapshot of current physician practice in US nursing homes. Such information is needed as the debate over the physician s role in the nursing home continues and new policy is framed that will ultimately define the future of medical practice in the nursing home. That 74% of the national survey respondents were members of AMDA suggests that the AMDA membership is representative of the national medical director population. (J Am Med Dir Assoc 2009; 10: 93 97) Keywords: Nursing homes; physician practice; medical director University of Rochester, Division of Geriatrics & Aging, Rochester, NY (T.V.C., J.K., P.R.K.). The authors have no conflicts of interest relating to this article. Address correspondence to Paul R. Katz, MD, University of Rochester School of Medicine and Dentistry, 435 East Henrietta, Rochester, NY 14620. E-mail: paul_katz@urmc.rochester.edu Copyright 2009 American Medical Directors Association DOI: 10.1016/j.jamda.2008.07.007 Nursing homes currently serve an increasingly frail population with complex medical needs. 1 Despite these trends, little is known regarding physician practice in this unique health care setting. 2 While a previous national survey indicated that only 23% of physicians see nursing home patients, and most spend 2 hours a month or less, 3 there are no existing national data that fully characterize the demographics of physicians caring for nursing home residents, the physician workloads, or the physician productivity and efficiency. Past and current ORIGINAL STUDIES Caprio et al 93
efforts to define resource utilization in nursing homes have included estimates of nurse and other professional staff time spent in patient care, but specifically excluded physician time. 4 Physician staffing estimates derived from the OSCAR (On-Line Survey Certification and Reporting) database do not reflect true physician staffing patterns. 5 Importantly, the relationship between physician practice and quality of care remains largely unexplored. Although nursing home physicians have been characterized as missing in action, their true role and impact on care remains elusive in the literature. 6 To begin to define more clearly physician practice in the nursing home, we administered a survey to a national randomly selected sample of nursing home medical directors. The purpose of the survey was threefold. First, it describes the medical practice, including actual time spent for routine encounters with nursing home residents and demographic characteristics of the physicians who serve as medical directors. Second, it determines the congruence between actual time spent for routine encounters with nursing home residents and the physician s view of the optimal or ideal time. The operative hypothesis is that physicians perceive a need to spend additional time per nursing home resident encounter and that this perception would be greatest in those physicians most knowledgeable of existing standards of care in the nursing home. Finally, it identifies salient barriers to optimal care by the practicing physician, as perceived by the nursing home medical director. METHODS Respondents Survey participants were selected from a listing of all Medicare-certified nursing facilities in the United States. 7 Using a table of random numbers, 200 medical director subjects were selected to participate in the survey. It was thought that a sound scientifically valid sample of nursing home medical directors, including both members and nonmembers of the American Medical Directors Association, would enhance generalizability of the results. A total of 100 surveys were returned for a response rate of 50%. Procedures The Dillman Total Design mail survey methodology was used. 8 An initial mailing consisted of the survey, an introductory letter, and a self-addressed return envelope. A unique tracking number identified each survey. One week later, a thank you/reminder postcard was mailed to all respondents. Following the post card mailing, a second survey packet was mailed 10 days later to all respondents who did not reply. A third mailing was sent 2 weeks later to the still remaining nonresponders. Finally, 2 weeks after the third mailing, a physician member of the research team (P.R.K.) called each of the nonresponding physicians to enlist their support of the study. To ensure confidentiality, the demographic information requested could not be used to identify any individual physician or facility. Completion of the survey constituted consent. Finally, survey results were converted into a computerized database and stored in an encrypted data file. The University of Rochester institutional review board approved this research methodology. Survey The survey items were culled from previous surveys 9,10 and pilot tested among medical directors practicing under the auspices of the Division of Geriatrics/Aging at the University of Rochester. The survey consisted of open- and closed-ended items that focused on the demographic characteristics of the medical director, demographic characteristics of the nursing home, and the extent of the medical director s nursing home practice, including the perceived ideal and reported actual time spent in nursing home visits for 4 common types of visits, ie, new admission for long-term care, new admission for rehabilitation, 30/60-day review, and readmission to the nursing home. Finally, a list of potential barriers to providing optimum visits in the nursing home was presented and the respondents were asked to rate how important each barrier was on a 5-point unipolar scale, anchored by not at all and very much. Mindful that medical directors could practice in several nursing homes, the medical directors were instructed to answer based on the facility that they considered to be their primary nursing home practice. RESULTS Nursing Home Characteristics As can be seen in Table 1, the medical director s primary facility had a mean size of 157 beds, was free standing, and was primarily proprietary. A mean of 16 physicians were credentialed to provide primary care in the nursing home. The nursing home medical staff organization was characterized as closed staff in one quarter of the nursing homes (as defined by no more than 10% of the nursing home residents seen by a community-based physician who was not employed by the nursing home). The medical directors had a formal agreement with a group of physicians to provide primary care in the nursing home in 17% of the homes. Two thirds of medical directors reported working with nurse practitioners or physician assistants (69%), with an average of 2.6 nonphysician providers per facility. Only a minority of nonphysician providers were employed by the nursing home (16%). Table 1. Nursing Home Characteristics of Respondents Principal Nursing Home Practice Mean number of beds 157 Sponsorship, % Proprietary 67 Voluntary (not for profit) 13 Voluntary (religiously affiliated) 14 Government 9 Facility part of nursing home chain, % 58 Physical location, % Free standing 91 On campus of hospital 8 Within hospital 1 Facility Joint Commission Accreditation, % 44 Facility provides subacute care, % 70 94 Caprio et al JAMDA February 2009
Table 2. Medical Director Demographics Median year graduated from medical school 1975 Mean years in long-term practice 18 Sex, % Male 83.5 Female 16.5 Discipline, % Family Medicine 37.0 Internal Medicine 51.0 General Practice 2.5 Psychiatry 1.0 Other 9.0 Completed Fellowship in Geriatrics, % 15.0 Certificate of Added Qualifications in Geriatrics, % 43.0 Certified Medical Director, % 41.0 American Medical Directors Association member, % 74.0 Faculty appointment in medical school, % 38.5 Medical Director Characteristics As can be seen in Table 2, the medical director respondents were most likely to be internal medicine (51%) or family medicine (37%) physicians, the majorities of whom were male and had been practicing in long-term care for more than 18 years. These physicians also devoted a sizeable proportion of their practice (44%) to nursing home care. While a large number (42%) had a certificate of added qualification (CAQ) in geriatrics, few had formal fellowship training (15%). Nearly three quarters of the respondents (74%) were members of the American Medical Directors Association (AMDA), 41% were certified medical directors (CMD), and over one third had faculty appointments. Nursing Home Practice As seen in Table 3, respondents, typically, were medical directors for 2 nursing homes. On average, they provided primary care in 4 homes (median 3). The medical directors reported spending a mean of 12 hours per week in the primary nursing home where they served as medical director. Respondents reported spending a mean of 31 hours per month in the nursing home working with nursing staff (median of 12 hours; range 1 240). When asked, relative to other patients with similar conditions, the medical directors reported that nursing home residents require an average of 33% more encounter time (median 25%; range 80 to 125) compared with patients in other practice settings. Perceptions of Actual and Optimum Visit Times Actual and optimum visit times (in minutes) are reported in Table 4 for 4 routine types of visits, and t-tests between the actual and perceived optimal visit times were computed. As seen in Table 4, overall, physicians reported significantly less time spent in visits for new admission for long-term care, new admission for rehabilitation, 30/60 day, and readmission to nursing home compared with what they judged was the optimal ideal time. Next, whether physician training and certification were related to the degree of incongruence between the actual and optimal visit times was examined. Difference scores were computed between the actual visit time reported and the judged optimal visit time for each of the 4 routine visits. Separate 1-way analyses of variance were performed that compared the difference scores between physicians with and without CMD. This was repeated comparing physician members of AMDA and those non-amda members, geriatric fellowship trained physicians and those who were not, and physicians with and without a certificate of added qualifications (CAQ) in geriatric medicine. Whether the physician had a CMD and whether the physician was a member of AMDA were not related to the incongruence between the actual and ideal visit times for any of the routine visits. However, a significant difference (F (1, 69) 15.47, P.001) emerged between physicians with and without geriatric fellowship training for the 30/60-day review visits. As can be seen in Table 4, fellowship-trained physicians reported greater discrepancies between the actual time spent in 30/60-day reviews and the judged optimal time compared with physicians who were not fellowship trained. The same pattern was found (F (1, 69) 12.16, P.001) when comparing physicians with CAQ (actual time X 19; optimal time X 26) to physicians without CAQ (actual time X 19; optimal time X 20) for the 30/60-day reviews. Again, for readmission evaluation visits a significant difference (F (1, 70) 6.67, P.01) emerged between physicians with and without fellowship training. As can be seen in Table 4, fellowshiptrained physicians reported greater discrepancies between the actual time spent for readmission revaluations of nursing home residents and the judged optimal time, compared with physicians who were not fellowship trained. The same significantly greater discrepancy was found (F (1, 70) 4.52, P.05) when comparing physicians with CAQ (actual time X 32; optimal time X 37) to physicians without CAQ (actual time X 30; optimal time X 31) for nursing home readmissions. Perceived Barriers to Providing Optimum Visits Barriers to providing optimum visits in the nursing home are presented in Table 5. Nursing support and accurate/accessible information were recorded as most problematic and reimbursement the least problematic barrier. As seen in Table 6, focusing on a sentinel event, treatment of uncomplicated pneumonia, lack of nursing staff support and nurse training Table 3. Directors Nursing Home Practice Characteristics of Medical Median number of facilities served as medical 2 director Median number of facilities provided primary care 3 Mean number of nursing homes practicing as 4.29 primary care physician Median number of nursing homes practicing as 3 primary care physician Percentage of practice devoted to nursing home 44 care Mean hours spent per week in nursing home 12 Mean hours functioning as medical director 2.3 ORIGINAL STUDIES Caprio et al 95
Table 4. Mean Medical Director Ratings of Actual Time Spent versus Ideal Times for Nursing Home Visits (in minutes) All Physicians Physicians With Fellowship Physicians Without Fellowship Type of Visit Type of Visit Type of Visit Actual Ideal Actual Ideal Actual Ideal New admission for long-term care 48 50* 55 62 47 49 New admission for rehabilitation 47 49 52 56 45 49 30/60 day review 19 22 17 28 19 21 Readmission to nursing home 32 33 34 43 31 32 * Significant difference comparing actual and ideal visit times, t (74) 3.15, P.02. Significant difference comparing actual and ideal visit times, t (72) 3.30, P.001. Significant difference comparing actual and ideal visit times, t (71) 4.19, P.001. Significant difference comparing actual and ideal visit times, t (73) 0.246, P.02. were cited as the reasons why not all nursing home residents with uncomplicated pneumonia could be treated in the nursing home. Lack of physician time and equipment were also noted as important factors in the decision to treat on site. DISCUSSION This study reports results from one of the first surveys of a nationally representative sample of nursing home medical directors that characterizes the major elements of physician practice in the nursing home. While the total sample is relatively small, it is generalizable given the random nature of the survey sample as well as a response rate of 50%, which is considered a very good return for physician respondents. 8 The fact that 74% of the survey participants were AMDA members speaks to the high penetration of AMDA nationally in the medical director community. Further, since our sampling frame was based on the listing of all Medicare-certified nursing facilities in the United States, the fact that three quarters of the respondents turned out to be AMDA members, provides empirical evidence that the AMDA membership is representative of the US medical director population. The additional nursing home and medical director demographics reported in this survey closely reflect other recent surveys. 5,9,10 The present survey indicates several extant barriers to optimal physician practice in the nursing home. Not surprisingly, the lack of accurate and accessible information was ranked as the most important barrier to optimum care in the nursing home. While these issues are recognized throughout the health care continuum, they are likely to be more problematic in the nursing home setting where electronic medical records are not as prevalent and where record keeping is less consistent from one facility to another. Lack of nursing support remained an important factor regarding the provision of care and in particular with the care for an acute illness (ie, pneumonia) in the nursing home. The fact that reimbursement issues were rated the least problematic barriers, although still in the moderate importance range, suggests that enhancing the quality of nursing home care may not simply reduce reimbursement issues. Interestingly, all of the 6 potential barriers presented in the survey scored above moderate importance, reinforcing the notion that care provision in nursing homes is complex and dependent on a number of related but distinct variables. Further research is needed that would test interventions designed to address specific barriers and assess the impact on quality markers. The medical director s perceptions of time necessary to complete routine nursing home visits mirrors the time allotment recommended by the Centers for Medicare and Medicaid Services. 11 For example, a 30- or 60-day note for a typical nursing home resident with multiple comorbidities would likely merit a CPT code of 99308 or 99309, corresponding to average physician times of 15 to 25 minutes versus respondents average of almost 19 minutes. Likewise, new admissions (rehabilitation presumed to be more complex) would warrant CPT codes of 99303 or 99306 for times between 35 and 45 minutes versus respondents average of 48 minutes. Readmissions to the nursing facility may qualify under any of the Table 5. Mean Medical Director Ratings of Perceived Barriers to Providing Optimum Visits in the Nursing Home Accessible information 4.74* Accurate information 4.68 Nursing support 4.53 Clinical support 3.60 Scheduling 3.26 Reimbursement 3.13 * The higher the number the more serious the perceived barrier, scale anchored by: 1 not at all; 2 slightly; 3 moderately; 4 quite a bit; 5 very much. Table 6. Mean Medical Director Ratings of Perceived Barriers to Treating Residents for Uncomplicated Pneumonias in the Nursing Home Lack of nursing support 3.22* Lack of nurse training 3.22 Lack of physician time 2.91 Lack of equipment 2.87 Lack of administration support 2.11 Reimbursement 1.92 * The higher the number the more serious the perceived barrier, scale anchored by: 1 not at all; 2 slightly; 3 moderately; 4 quite a bit; 5 very much. 96 Caprio et al JAMDA February 2009
above codes, depending on whether the resident had been formally discharged from the nursing home and/or had developed significant change in condition. The relationship between physician attributes (fellowship training; CAQ) and perceptions of time necessary for certain nursing home visits (30- to 60-day notes; readmission to nursing home) suggests that geriatric training impacts perceived standards of care. This is particularly relevant in the context of the recently released Institute of Medicine report Retooling for an Aging America: Building the Health Care Workforce, 12 which calls for more geriatric training at all levels as well as initiatives to increase recruitment and retention of geriatric providers. CONCLUSION The present study provides a snapshot of current physician practice in US nursing homes. Such information is needed as the debate over the physician s role in the nursing home continues, and new policy is framed that will ultimately define the future of medical practice in the nursing home. Additional studies will need to examine the relationship between medical director characteristics, practice patterns, cost, and clinical outcomes. Finally, the discrepancy between certified and noncertified geriatric physician-medical directors on how they see their practice relative to an ideal standard, suggests that physician expectations and definition of gold standards may be influenced by the degree of specialized geriatric training. REFERENCES 1. Kasper J, O Malley M. Changes in Characteristics, Needs and Payments for Elderly NH Residents: 1999 2004. June 2007. Available at: www. kff.org/medicaid/upload/7663.pdf. Accessed May 15, 2008. 2. Levy C, Palat SI, Kramer AM. Physician practice patterns in nursing homes. J Am Med Dir Assoc 2007;8:558 567. 3. Katz PR, Karuza J, Kolassa J, Hutson A. Medical practice with nursing home residents: Results from the National Physician Professional Activities Census. J Am Geriatr Soc 1997;45:911 917. 4. Cooney LM Jr, Fries BE. Validation and use of resource utilization groups as a case-mix measure for long-term care. Med Care 1985;23:123 132. 5. Feng Z, Katz PR, Intrator O, et al. Physician and nurse staffing in nursing homes: The role and limitations of the Online Survey Certification and Reporting (OSCAR) system. J Am Med Dir Assoc 2005;6:27 33. 6. Katz PR, Karuza J. Physician practice in the nursing home: Missing in action or misunderstood. J Am Geriatr Soc 2005;53:1826. 7. Directory of Nursing Homes. Phoenix, AZ: Oryx Press. 2005. 8. Dillman D. Mail and Telephone Surveys: The Total Design Method. New York: John Wiley and Sons, 1978. 9. Karuza J, Katz PR. Physician staffing pattern correlates of nursing home care: An initial inquiry and consideration of policy implications. J Am Geriatr Soc 1994;42:787 793. 10. Katz PR, Karuza J, Parker M, Tarnove L. A national survey of medical directors. Journal of Medical Direction 1992;2:81 94. 11. American Medical Directors Association. Descriptions of Long-Term Care CPT Codes. Available at: www.amda.com/members/advocacy/ cptcentral/codes.cfm. Accessed May 20, 2008. 12. Institute of Medicine. Retooling for an Aging America: Building the Health Care Workforce. Washington, DC: National Academy Press, 2008. ORIGINAL STUDIES Caprio et al 97