Primary Care By Leighton Ku, Bianca K. Frogner, Erika Steinmetz, and Patricia Pittman
Community Health Centers Employ Diverse Staffing Patterns, Which Can Provide Productivity Lessons For Medical Practices Community health centers are at the forefront of ambulatory care practices in their use of nonphysician clinicians and team-based primary care. We examined medical staffing patterns, the contributions of different types of staff to productivity, and the factors associated with staffing at community health centers across the United States. We identified four different staffing patterns: typical, high advanced-practice staff, high nursing staff, and high other medical staff. Overall, productivity per staff person was similar across the four staffing patterns. We found that physicians make the greatest contributions to productivity, but advanced-practice staff, nurses, and other medical staff also contribute. Patterns of community health center staffing are driven by numerous factors, including the concentration of clinicians in communities, nurse practitioner scope-of-practice laws, and patient characteristics such as insurance status. Our findings suggest that other group medical practices could incorporate more nonphysician staff without sacrificing productivity and thus profitability. However, the new staffing patterns that evolve may be affected by characteristics of the practice location or the types of patients served. ABSTRACT
N
umerous areas of the United States have an inadequate supply of primary care physicians, and the shortfall is expected to grow and spread.1 To address the shortage of primary care physicians and improve the quality of care delivered, health policy experts have recommended that medical practices transform themselves and embrace “team-based” care, encouraging a broader range of medical staff to work together.2–5 But to what extent should practices expand the use of nurse practitioners, physician assistants, nurses, or other medical staff, and what are the staffing implications? The paths to transforming primary care staffing and ambulatory care practices are not well mapped. Community health centers are federally
10.1377/hlthaff.2014.0098 HEALTH AFFAIRS 34, NO. 1 (2015): 95–103 ©2015 Project HOPE— The People-to-People Health Foundation, Inc.
doi:
Leighton Ku ([email protected]) is a professor in the Department of Health Policy, Milken School of Public Health, George Washington University (GWU), in Washington, D.C. Bianca K. Frogner is an associate professor in the Department of Family Medicine at the University of Washington, in Seattle. At the time of this research, she was an assistant professor in the Department of Health Services Management and Leadership, Milken School of Public Health, GWU. Erika Steinmetz is a senior research scientist in the Department of Health Policy, Milken School of Public Health, GWU. Patricia Pittman is an associate professor in the Department of Health Policy, Milken School of Public Health, GWU.
funded safety-net organizations that collectively provided comprehensive primary care to more than twenty-one million low-income patients in 2013.6 They have been in the vanguard of teambased care and flexibly vary the roles of medical staff based on local needs and staff capabilities. Community health centers use advanced-practice clinicians, such as nurse practitioners or physician assistants, more than regular private medical practices do.7 Finding and retaining physicians has long been a problem for community health centers, in part because they are located in medically underserved areas, which typically have shortages of primary care physicians.8 Community health centers have learned how to adjust their staffing while continuing to provide high-quality care in an affordable and efficient manner. CenJ a n u a ry 2 0 1 5
3 4: 1
Health Affai rs
95
Primary Care ter directors often say that they encourage staff to “work at the top of their capabilities,” taking greater responsibility for patient care consistent with their training, experience, and licensure, instead of working in rigidly defined roles. But staffing patterns can vary widely from one health center to another. An important factor in the extent to which staffing can be diversified is the effect of staffing patterns on productivity—a key measure of the efficiency of care. Higher productivity can reduce personnel costs or increase the number of visits, in either case increasing revenue from patients. Traditionally, productivity is measured based on the number of visits in which a physician (or advanced-practice staff member) sees a patient. This ignores the roles of other staff, however. A single clinician is not usually the sole provider of a patient’s care. A medical assistant may take vital signs; a physician may conduct the evaluation and make a diagnosis; and a nurse may draw blood, administer an injection, or educate the patient. According to a team-based care perspective, the visit should be viewed in terms of the joint productivity of the overall team. Instead of thinking of the visit as 100 percent produced by the physician, we might view 75 percent of the visit as produced by the physician, 10 percent by the medical assistant, and 15 percent by the nurse. Even staff who are not in the exam room play a role in medical productivity. For example, laboratory staff contribute to diagnoses, and quality assurance staff help monitor and improve care quality. Although they do not see patients, they are part of today’s medical practice system. This article examines different medical staffing patterns in community health centers, the impact of staffing on productivity, staff roles, and the factors that affect staffing patterns. The experiences of community health centers are relevant not only for safety-net providers, but also for the broader array of group medical practices in the United States as they implement team-based care approaches.
Study Data And Methods Data Our primary data source was the annual reports for 2012 filed with the Uniform Data System (UDS) by all community health centers that received federal section 330 grants, including sites in all fifty states, the District of Columbia, and US territories.6 Section 330 grants are the core federal grants provided to nonprofit community health centers by the Health Resources and Services Administration. All staff levels in these reports are expressed as full-time equivalents. The UDS also reports the 96
Health Affairs
J a n u a ry 2 0 1 5
3 4: 1
number of medical visits and patients (as opposed to dental or mental health visits and patients), diagnostic categories of visits, and other key caseload and financial characteristics of community health centers. There were 1,198 community health centers in the United States in 2012. We excluded seven whose reports were missing data, resulting in a final sample of 1,191 centers. Community health centers may operate clinics in multiple sites, and these 1,191 centers represented about 8,900 clinical sites in 2012. We classified medical staff into four categories: physicians (medical doctors and doctors of osteopathy; most were in primary care specialties), advanced-practice staff (nurse practitioners, physician assistants, and certified nurse-midwives), nurses (registered, licensed vocational, and practical nurses), and other medical staff (such as medical assistants, nurse aides, and lab or radiology staff). We excluded mental health, dental, vision, enabling service (for example, interpretation service), and administrative or finance visits and staff to focus on the core medical services available in most medical practices. We recognize the importance of mental health and dental care. However, most primary care practices do not employ behavioral health or dental staff.9 The UDS data also include the number of medical patients served at the health centers and the percentages of patients who are uninsured, are covered by Medicaid, and have limited English proficiency. We measured productivity as the number of weighted medical visits per center in 2012. Weighting is important because medical visits vary in the complexity of care required, which may also influence the type of staff involved. Diagnostic categories classified according to the International Classification of Diseases, Ninth Revision (ICD-9), codes were reported for about two-third of the UDS visits.We used data from the 2011 Medical Expenditure Panel Survey to compute average expenditures for office-based care for those diagnostic categories.We used the casemix of diagnoses and average expenditure per diagnosis to compute an expenditure index for each center. A center with a higher score tended to have more complex cases than a center with a low score (for more methodological information and analyses, see the online Appendix).10 By weighting visits, we sought to adjust for the complexity of different types of care required, depending on the diagnosis. To help understand workforce and policy environments that could affect staffing, we merged UDS data with county- and state-level data. We included county-level Area Health Resource File11
data about the concentrations of primary care physicians, advanced-practice staff, and nurses per 1,000 county residents and the percentage of the population in each county with incomes below the federal poverty level. We also included a rural-urban continuum code, ranging from 1 (the county was in a metropolitan area with a population of more than one million) to 9 (the most rural locale).12 We also merged the UDS data with state-level data on nurse practitioners’ scope of practice, or the range of services that nurse practitioners can conduct without a physician’s oversight in a given state. Nurse practitioners’ scopes of practice were categorized as full (nurse practitioners could treat patients or prescribe medications without a physician’s authorization), partial (nurse practitioners could treat but not prescribe), or restricted (most nurse practitioners’ actions must be authorized by a physician).13 Analysis To identify staffing patterns, cluster analysis was used. This is a method of pattern recognition that identifies clusters of similar observations—in this case, based on the percentages of staff in each of four categories in each health center. Each cluster was designed to differ maximally from the other clusters. We also examined the patterns of staffing by estimating correlation matrices. The UDS has limited information about who was present for each visit, showing only the type of lead clinician—usually a physician—who was responsible for the visit. To show the contributions of all medical staff and to compute joint productivity, we used the overall medical staff composition of the health center and number of weighted visits to compute the productivity of each staff type using multivariate ordinary least squares regression, with robust standard errors. This standard statistical approach, used in many related types of analyses, let us estimate the contributions of all medical staff, including some who might not see patients but who nonetheless contribute to the overall medical enterprise. Limitations This study was limited by the type and quality of data available. We could measure the number and types of staff at the community health centers. However, we could not measure their actual roles in patient care or how staff interacted to form effective teams. Further research is needed to obtain a more finely grained view of the roles and responsibilities of different types of health professionals working in teams. In addition, this study focused on the role of medical staff and did not include information about behavioral, dental, vision, enabling service, or administrative staff, who also have important roles to play. There may be some misclassification in staff
roles in the UDS. For example, a nurse employed at a community health center may serve in an administrative capacity or staff a health program that does not contribute to the reported number of medical visits (such as a Women, Infants, and Children nutrition program or a health education program). The UDS instructions ask that staff be classified by their main job function, but staff often serve multiple functions. A final limitation is that the approach we used to weight medical visits might not adequately reflect real differences in the amounts and types of resources across community health centers. However, comparisons with unweighted models suggested that weighting increased the effect of physicians, which is consistent with evidence that physicians tend to treat more complex cases than other staff do.
Study Results Staff Composition The median number of fulltime-equivalent medical staff at a community health center was 27.2, while the mean was 44.6. The median number of physicians at a health center was 5.0, and the mean was 8.8. This is comparable to the typical medical practice for American physicians: The median physician works in a group practice of 5–9 doctors.14 On average, physicians made up 19.0 percent of the medical staff at community health centers (Exhibits 1 and 2). But the size and staffing composition varied widely, as shown below. Patterns Of Staffing Using cluster analysis, we identified four dominant staffing patterns, Exhibit 1 Composition Of Medical Staff In Community Health Centers Overall And For Four Staffing Clusters, 2012
SOURCE Authors’ analysis of 2012 data from the Uniform Data System. NOTE Sample sizes are in Exhibit 2.
J a n u a ry 201 5
34:1
Health A ffairs
97
Primary Care Exhibit 2 Staffing And Other Characteristics, By Community Health Center (CHC) Staffing Clusters, 2012 Type of cluster Characteristic Number of CHCs
All CHCs 1,191
Typical 421
High advancedpractice staff 44
High nursing staff 295
High other medical staff 431
Mean percent of staff Physicians Advanced-practice Nurses Other medical
19.0% 17.9 25.6 37.4
19.1% 15.2 27.5 38.2
11.7% 54.4 19.2 14.7
21.2% 17.4 48.9 12.5
18.2% 17.2 8.6 56.1
44.6 2.58 15,137 50,234 1,123 52.1%
52.5 2.21 17,179 56,898 1,071 52.3%
10.4 0.24 5,161 13,545 1,261 38.6%
30.1 4.67 11,009 37,516 1,223 32.9%
50.1 1.85 16,985 56,192 1,091 66.6%
0.4% 38.7 34.5
15.0% 36.7 36.0
15.1% 46.9 28.6
14.2% 35.6 31.6
23.0% 41.9 35.4
Mean FTE medical staff Mean physicians per advanced-practice staff Medical patients Weighted medical visits Mean weighted visits per FTE medical staff Percent urban centers Mean percent of patients with: Limited English proficiency No insurance Medicaid
SOURCE Authors’ analysis of 2012 data from the Uniform Data System. NOTES All differences across the four groups are significant (p < 0:001). For details about the types of staff, see the “Data” section. FTE is full-time equivalent.
based on the percentages of medical staff in the four categories. The patterns were typical, high advanced-practice staff, high nursing staff, and high other medical staff. The typical group had a staff distribution similar to the overall national average (Exhibits 1 and 2). In the high advancedpractice staff group, about half of the staff were advanced-practice clinicians. Similarly, in the high nursing staff group, almost half of the staff were nurses, and in the high other medical staff group, about half of the staff were other medical professionals. The high advanced-practice staff group had a lower percentage of physicians than any other group. Across all centers, there were 2.58 physicians for every advanced-practice staff member (Exhibit 2). But in centers with high advancedpractice staff, the advanced-practice staff outnumbered physicians by about three to one. In high nursing staff centers, there were 4.67 physicians per advanced-practice staff member. The percentages of nurses and other medical staff at a community health center were strongly negatively correlated (r ¼ −0:84), meaning that centers with more nurses had fewer other medical staff. The percentages of physicians and advanced-practice staff were also negatively correlated (r ¼ −0:52), meaning that centers with more physicians had fewer advanced-practice staff (for the full correlation matrix, see the Appendix).10 The four clusters varied in a number of ways (Exhibit 2). Health centers in the typical and 98
H ea lt h A f fai r s
January 2015
34:1
high other medical staff clusters tended to be the largest, with roughly fifty medical staff, 17,000 medical patients, and 56,000 weighted medical visits, on average. High advanced-practice staff centers were the smallest. A basic measure of productivity is the number of weighted medical visits per staff member, without regard to staff type. The average levels were similar across the clusters, ranging from 1,071 to 1,261 visits per staff member (Exhibit 2). The levels were slightly higher for the high advanced-practice staff and high nursing staff centers. We did not have average salary information by category, so we could not determine the relative labor cost per visit. There may be a variation in the types of care provided in the different clusters, but it is important to remember that visits were weighted to adjust for differences in patients’ diagnoses. Of the four clusters, centers with high other medical staff were the most likely to be urban (66.6 percent) and to have the greatest share of patients with limited English proficiency (23.0 percent; Exhibit 2). Centers with high nursing staff were the most likely to be rural. Centers with high advanced-practice staff had the highest proportion of uninsured patients but the lowest percentage of patients with Medicaid. Contributions To Productivity Regression models were used to measure the contributions of different types of staff to overall productivity
(Exhibit 3). Not surprisingly, physicians consistently had the highest marginal contribution to productivity. Overall, one additional physician was associated with 2,994 additional visits. The marginal contribution of nurses appeared to be the lowest (292 additional visits), but the coefficient in that case was not significant. Because we were measuring joint productivity, the results were different than if we had used traditional measures of productivity. For example, there would have been 3,351 visits per physician if any physician-led visit was counted as belonging only to the physician, instead of the 2,994 visits estimated by our model, which assumed that there were also contributions by other staff. The marginal contributions model was also run separately for each of the four clusters. The results differed somewhat across cluster types. In high advanced-practice centers, advanced practitioners had higher contributions. Similarly, the effects of nurses and other medical staff were larger in high nurse or high other medical centers, respectively. For example, in the high nursing staff centers, an additional nurse was associated with 1,407 additional visits (Exhibit 3). In some models, the coefficients for nurses or for other medical staff were not significant. This might have occurred because the levels of nurses and of other medical staff were negatively correlated, so they canceled each other out. As a check, we developed models that pooled
nurses and other medical staff. In these consolidated models, the coefficients for the combined nurses and other medical staff were positive and significant overall and in all but one of the cluster-specific models. These results support the notion that these staff members contribute to productivity, but their negative correlation leads them to cancel each other out when both are in the models. As described in the Appendix,10 we tested alternative models using unweighted visits as the productivity metric. The results were generally similar. However, the weighted models’ coefficients for physicians tended to be higher, while those for advanced-practice staff were lower. This signals that weighting increases the apparent contribution of physicians and decreases the apparent contribution of advanced-practice staff, which suggests that physicians are more involved in complex visits, compared to the advanced-practice staff. Factors Affecting Staff Composition We also examined the factors that might explain why community health centers have varying staffing patterns. Exhibit 4 presents the results of regression models that examined the relationship of these factors to the percentages of medical staff who were physicians, advanced-practice staff, nurses, and other medical staff. We examined the roles of factors related to clinician availability, state policies, and health center caseloads. Rural or urban location was critical in all four models. The analyses show that rural location
Exhibit 3 Contributions To Medical Visit Productivity Per Staff Person, By Type Of Community Health Center (CHC) Staffing Cluster, 2012 Type of cluster
Physicians
Advancedpractice staff
Nursing staff
Other medical staff
Adjusted R2
All Coefficient 95% CI
2,994**** (1,955, 4,036)
1,584**** (1,394, 2,073)
292 (−119, 704)
548**** (352, 745)
0.952 —a
Typical Coefficient 3,370**** 95% CI (2,746, 3,995) High advanced-practice staff
1,546**** (907, 2,185)
347 (−241, 936)
265 (−115, 645)
0.963 —a
Coefficient 95% CI
2,761**** (1,877, 3,646)
2,287*** (974, 3,600)
4 (−1,159, 1,166)
−727 (−2,158, 703)
0.938 —a
2,086**** (1,317, 2,854)
198 (−734, 1,138)
1,407**** (922, 1,892)
357 (−239, 953)
0.948 —a
1,664**** (761, 2,567)
−788 (−1,837, 259)
744**** (322, 1,165)
0.945 —a
High nursing staff Coefficient 95% CI
High other medical staff Coefficient 2,923**** 95% CI (1,159, 4,686)
SOURCE Authors’ analysis of 2012 data from the Uniform Data System. NOTES The coefficients estimated by the models indicate the effect of adding one staff member on the number of weighted medical visits. For example, one additional advanced-practice staff member was associated with 1,584 additional visits. For details about the types of staff, see the “Data” section. CI is confidence interval. aNot applicable. ***p < 0:01 ****p < 0:001
January 2015
34:1
H e a lt h A f fai r s
99
Primary Care Exhibit 4 Coefficients For Factors Associated With Percentage Of Staff In Community Health Centers (CHCs), 2012 Dependent variable—percent of staff who are: Factor County characteristics
Physicians
Nursing staff
Other medical staff
−0.63****
0.76****
3.16**** −3.21**** 0.36*** 0.01
−3.46**** 4.86**** −0.54*** −0.14**
−2.09*** −0.38
2.02** 0.37
−7.80**** −0.35
7.87**** 0.35
0.01 −10.92**** −2.16 0.02
−0.10**** 10.96**** 0.88 −0.05***
0.01 −3.48 −6.44 −0.18****
0.08** 3.44 7.72 0.21****
24.30**** 0.109
15.95**** 0.151
25.49**** 0.185
34.26**** 0.204
Rural-urban continuum Per 1,000 population Physicians Advanced-practice staff Nurses Percent of population in poverty State scope-of-practice laws Full versus restricted scope Partial versus restricted scope CHC patients Number (1,000s) Percent uninsured Percent with Medicaid Percent with limited English proficiency
Advancedpractice staff
Constant Adjusted R2
2.13*** −0.08 1.31 0.43 −0.09
−2.26**** 0.38 −2.95** −0.24 0.22**
SOURCE Authors’ analysis of 2012 data from the Uniform Data System. NOTE For details about the types of staff, rural-urban continuum, and scope of practice laws, see the “Data” section. **p < 0:05 ***p < 0:01 ****p < 0:001
was associated with more advanced-practice staff and nurses, and with fewer physicians and other medical staff, being employed at a health center. Staffing was affected by the availability of health care professionals in the area. Community health centers in counties with more primary care physicians per 1,000 population had more physicians and fewer advanced-practice staff (Exhibit 3). Similarly, centers in counties with more advanced-practice staff per 1,000 population had more of that type of staff member and fewer physicians. Interestingly, centers in counties with higher concentrations of advancedpractice staff also had fewer other medical staff. A higher county-level concentration of nurses was associated with more physicians and fewer advanced-practice staff at centers but did not significantly affect the share of nurses on the staff. Nurse practitioner scope-of-practice laws were also associated with community health center staffing. We examined the effects of full and partial scopes of practice, compared to restricted scope (Exhibit 4). Centers in states with full scope of practice used slightly fewer physicians and slightly more advanced-practice staff. Full scope of practice was also associated with fewer non-advanced-practice nurses, but more other medical staff. Finally, we examined the characteristics of community health center caseloads (Exhibit 4). 100
Health Affairs
J a n u a ry 201 5
34:1
Other medical staff were more common in centers with more patients, while the opposite was true for advanced-practice staff. Community health centers with more uninsured patients had fewer physicians and more advanced-practice staff, but the percentage of patients with Medicaid was not significantly associated with staffing differences. The share of patients with limited English proficiency was associated with greater use of other medical staff.
Discussion Our analyses indicate that physicians are the leading contributors to productivity, but other medical staff also make meaningful contributions and increase the amount of care that can be delivered. Given concerns about primary care physician shortages, our findings suggest that greater use of nonphysician medical staff could enable more patients to be seen with limited resources. The average ratio of physicians to nonphysician medical staff for community health centers nationwide was 1:4. In comparison, a study of practices in a comprehensive primary care demonstration found a ratio of 1:2.5, which is still a higher ratio than that found in other practices.9 The ability to implement effective team-based primary care is relevant not only to the future of community health centers, but also to the broader system of group medical prac-
There is further room for innovation in the effective use of lowerlevel staff.
tice in the United States. We found diverse staffing patterns in community health centers, which suggests that staff roles and functions vary across settings. Across the four common staffing patterns that we identified, the overall levels of productivity per staff were relatively similar, even after we adjusted for the diagnostic complexity of the visits. There did not appear to be a distinctly optimal mix of staffing among the four patterns: Practices can be productive with varying blends of physicians, advanced-practice staff, nurses, and other medical staff. We found that local factors—such as rural location, the availability of clinicians in the area, scope-of-practice laws, and the patient and revenue mix of the centers—may be the most important determinants of staff composition. The data do not directly tell us what roles each type of staff play, but we can infer broad patterns. Nurses and other medical staff typically function as complements to physicians or advanced-practice staff, helping them reduce the time they spend on tasks that require less specialized skills. However, the strong negative correlation between nurses and other medical staff shows that more use of nurses is associated with less use of other medical staff, and vice versa. This result suggests that their roles may be somewhat interchangeable.6 In addition, nurses and other medical staff appear to make lower contributions to productivity than physicians or advanced-practice staff do. It is plausible that many of the functions that nurses and other medical staff fulfill—including patient education, quality assurance, and scheduling and administrative tasks—are necessary for care but not as strongly correlated with an increased number of medical visits, compared to the typical activities of physicians. Thus, nurses and other medical staff appear to have lower productivity. There is further room for innovation in the effective use of lower-level staff. For example, a recent study showed that the use of medical assistants as health coaches improved patient outcomes.15 More detailed research, including qualitative studies, could augment our under-
standing of the roles of nurses and other medical staff in primary care. We found a moderate negative correlation between physicians and advanced-practice clinicians, which also indicates some interchangeability of roles. In some cases, physicians and advanced-practice staff serve similar functions as primary care providers—for instance, both can make diagnoses or issue prescriptions— and thus can serve as substitutes for each other. Even in states with restrictive scope-of-practice laws, advanced-practice staff are able to diagnose and prescribe, with a physician’s authorization. In other cases, advanced-practice staff may serve as complements to physicians. For instance, a physician may see patients with more complex conditions or initiate care for a patient, while advanced-practice staff provide important follow-up care or patient education.2,9,16,17 A recent analysis of data from the National Ambulatory Medical Care Survey examined the use of physicians and advanced-practice staff in community health centers and found that advancedpractice staff were more involved than physicians with preventive patient education and mental health care visits.18 Our analyses support evidence that physicians tend to be more involved in care for patients with complex conditions, compared to nonphysician medical staff. Staffing patterns differ across community health centers, which indicates that staff functions also differ. But in general, regardless of the staffing mix, the centers’ overall productivity is relatively similar. Instead of having a one-sizefits-all approach to staffing and care provision, community health centers seem to be flexible and take advantage of the staff they have. Diverse staffing patterns are also related to geographic differences, such as rural or urban location, and the concentration of primary care physicians or advanced-practice staff in the area. Community health centers are located in areas that are designated by the Health Resources and Services Administration as Medically Underserved Areas, which typically have a low concentration of physicians. But physician workforce shortages affect other types of practices across the country, and the issue is likely to become even more pressing in the coming years. We also found other factors that affected staff distribution. Community health centers with more advanced-practice staff had more uninsured patients. A possible explanation is that centers with more uninsured patients are under financial strain that makes it harder for them to attract or retain physicians, and thus more advanced-practice staff are used. In addition, we found that community health centers with high levels of other medical staff had more patients January 2015
34:1
Health A ffairs
101
Primary Care with limited English proficiency, which suggests that these staff are providing additional services such as language interpretation. The distribution of staff differed in states where nurse practitioners had full scope of practice. Community health centers in these states had more advanced-practice staff and other medical staff, but fewer physicians and nurses. This is not necessarily a causal relationship: The adoption of scope-of-practice laws is not random, and it is possible that the relationship occurs because states with fewer physicians adopted full scope of practice to compensate. Regardless, it is reasonable to believe that granting nurse practitioners full scope of practice gives greater flexibility to community health centers and other medical practices in staffing choices, which is helpful if primary care physicians are in short supply. An Institute of Medicine report recommended expanding nurse practitioners’ scope of practice.19 We found that the use of advanced-practice staff was greatest in smaller community health centers that were more rural and had more uninsured patients. Previous studies have found that advanced-practice professionals can offer high-quality primary care services, often comparable to those provided by physicians.16,20,21 It is beyond the scope of this study to assess whether some staffing patterns are better or worse than others or lead to better patient outcomes or greater patient satisfaction. Other research has found that in general, community health centers provide high-quality primary care, broadly comparable to that offered in other ambulatory care settings, despite the resource limitations that health centers face and the low incomes and spotty insurance coverage of their patients.22–25 Community health centers have been in the vanguard of innovations in medical practice. As of 2013, 80 percent of community health centers had electronic health record systems, and 54 percent were recognized as patientcentered medical homes.26 The number of community health centers in the United States has grown rapidly for more than a decade and is expected to continue to grow. It seems likely that the centers will continue to innovate and be flexible in staffing and defining staff roles, while also providing highquality care.
Conclusion As the number of community health centers grows, leading to potential staffing pressures, centers must develop innovative staffing and
102
Health Affairs
J a n u a ry 201 5
34:1
The distribution of staff differed in states where nurse practitioners had full scope of practice.
care approaches to provide high-quality primary care in an efficient manner. The examples of community health centers may provide insights for other medical practices. In the broader US health care system, group medical practices are now facing many of the same pressures that community health centers have long encountered. There has been substantial research (and sometimes debate) about the broader roles of physicians and advanced-practice staff. However, there has been less attention to the important ways in which nurses and other medical staff contribute to care. Our findings indicate that it is possible to include a broad array of advanced-practice staff, nurses, and other medical staff in ambulatory medical practices to work alongside physicians and make meaningful contributions to productivity. Increasing the share of nonphysician staff may be more feasible for larger group practices than for small or solo practices because the larger size makes it possible to deploy staff in more diverse ways. Multiple staffing patterns, using more nonphysician staff in team-based care, are possible and could increase the number of patients served. In addition to regulatory issues such as scope of practice, insurers may need to consider whether their payment or credentialing policies are creating barriers to more flexible use of medical staff in patient care. For example, even if a state has full scope of practice for nurse practitioners, there may be problems if an insurer will pay only a physician for patient care services. We found that productivity levels were relatively similar across all staff configurations and that there did not appear to be a single optimal model of staffing. Both community health centers and the broader set of ambulatory practices can use diverse approaches that fit their needs and capacities and those of their communities. ▪
This research was presented at the AcademyHealth Annual Research Meeting in San Diego, California, June 10, 2014. This article was supported in part under a cooperative
agreement with the National Center for Health Workforce Analysis, Health Resources and Services Administration. The Uniform Data Systems data used in this article were provided by the Office
of Data and Quality, Bureau of Primary Health Care, Health Resources and Services Administration.
NOTES 1 Petterson SM, Liaw WR, Phillips RL, Rabin DL, Meyers DS, Bazemore AW. Projecting US primary care workforce needs: 2010–2025. Ann Fam Med. 2012:10(6):503–9. 2 Blumenthal D, Abrams MK. Putting aside preconceptions—time for dialogue among primary care clinicians. N Engl J Med. 2013;368(20):1933–4. 3 Chen EH, Bodenheimer T. Improving population health through teambased panel management: comment on “Electronic medical record reminders and panel management to improve primary care of elderly patients.” Arch Intern Med. 2011; 171(17):1558–9. 4 Bodenheimer TS, Smith MD. Primary care: proposed solutions to the physician shortage without training more physicians. Health Aff (Millwood). 2013;32(11):1881–6. 5 Auerbach DI, Chen PG, Friedberg MW, Reid R, Lau C, Buerhaus PI, et al. Nurse-managed health centers and patient-centered medical homes could mitigate expected primary care physician shortage. Health Aff (Millwood). 2013;32(11):1933–41. 6 Health Resources and Services Administration. Health center data [Internet]. Rockville (MD): HRSA; [cited 2014 Nov 17]. Available from: http://bphc.hrsa.gov/healthcenter datastatistics/ 7 Hing E, Hooker RS, Ashman JJ. Primary health care in community health centers and comparison with office-based practice. J Community Health. 2011;36(3):406–13. 8 Rosenblatt RA, Andrilla CH, Curtin T, Hart LG. Shortages of medical personnel at community health centers: implications for planned expansion. JAMA. 2006;295(9): 1042–9. 9 Peikes DN, Reid RJ, Day TJ, Cornwell DD, Dale SB, Baron RJ, et al. Staffing patterns of primary care practices in the Comprehensive Primary Care Initiative. Ann Fam Med. 2014;12(2):142–9. 10 To access the Appendix, click on the
11
12
13
14
15
16
17
18
Appendix link in the box to the right of the article online. Health Resources and Services Administration. Area Health Resources File [Internet]. Rockville (MD): HRSA; [cited 2014 Nov 24]. Available from: http://ahrf.hrsa.gov Economic Research Service. Rural/ urban continuum codes [Internet]. Washington (DC): Department of Agriculture; [last updated 2013 May 10; cited 2014 Nov 24]. Available from: http://www.ers.usda.gov/ data-products/rural-urbancontinuum-codes.aspx Pearson L. The 2012 Pearson report: a national overview of nurse practitioner legislation and health care issues. Monroe (NJ): NP Communications LLC; 2012. Kane CK, Emmons DW. New data on physician practice arrangements: private practice remains strong despite shifts in hospital employment [Internet]. Chicago (IL): American Medical Association; c2013 [cited 2014 Nov 12]. (Policy Research Perspectives). Available from: http:// www.nmms.org/sites/default/files/ images/2013_9_23_ama_survey_ prp-physician-practicearrangements.pdf Willard-Grace R. Not just a medical assistant anymore: new healthcare delivery models that use unlicensed professionals. Paper presented at: AcademyHealth Annual Research Meeting; 2014 Jun 9; San Diego, CA. Donelan K, DesRoches CM, Dittus RS, Buerhaus P. Perspectives of physicians and nurse practitioners on primary care practice. N Engl J Med. 2013;368(20):1898–906. Everett C, Thorpe C, Palta M, Carayon P, Bartels C, Smith MA. Physician assistants and nurse practitioners perform effective roles on teams caring for Medicare patients with diabetes. Health Aff (Millwood). 2013;32(11):1942–8. Downie S. Comparison of primary care visits to physicians and nonphysician providers in community
19
20
21
22
23
24
25
26
health centers and hospital outpatient departments. Paper presented at: AcademyHealth Annual Research Meeting; 2014 Jun 10; San Diego, CA. Institute of Medicine. The future of nursing: leading change, advancing health. Washington (DC): National Academies Press; 2011. Mundinger MO, Kane RL, Lenz ER, Totten AM, Tsai WY, Cleary PD, et al. Primary care outcomes in patients treated by nurse practitioners or physicians: a randomized trial. JAMA. 2000:283(1):59–68. Stanik-Hutt J, Newhouse RP, White KM, Johantgen M, Bass EB, Zangaro G, et al. The quality and effectiveness of care provided by nurse practitioners. J Nurse Pract. 2013;9(8): 492–500. Goldman LE, Chu PW, Tran H, Romano MJ, Stafford RS. Federally qualified health centers and private practice performance on ambulatory care measures. Am J Prev Med. 2012;43(2):142–9. Bruen BK, Ku L, Lu X, Shin P. No evidence that primary care physicians offer less care to Medicaid, community health center, or uninsured patients. Health Aff (Millwood). 2013;32(9):1624–30. Probst JC, Laditka JN, Laditka SB. Association between community health center and rural health clinic presence and county-level hospitalization rates for ambulatory care sensitive locations: an analysis across eight US states. BMC Health Serv Res. 2009;9:134. Richard P, Ku L, Dor A, Tan E, Shin P, Rosenbaum S. Cost savings associated with the use of community health centers. J Ambul Care Manage. 2012;35(1):50–9. Health Resources and Services Administration. 2013 health center data [Internet]. Rockville (MD): HRSA [cited 2014 Nov 12]. Available from: http://bphc.hrsa.gov/uds/ datacenter.aspx?year=2013
January 2015
34:1
Health A ffairs
103
Community Health Centers Employ Diverse Staffing Patterns, Which Can Provide Productivity Lessons For Medical Practices Community health centers are at the forefront of ambulatory care practices in their use of nonphysician clinicians and team-based primary care. We examined medical staffing patterns, the contributions of different types of staff to productivity, and the factors associated with staffing at community health centers across the United States. We identified four different staffing patterns: typical, high advanced-practice staff, high nursing staff, and high other medical staff. Overall, productivity per staff person was similar across the four staffing patterns. We found that physicians make the greatest contributions to productivity, but advanced-practice staff, nurses, and other medical staff also contribute. Patterns of community health center staffing are driven by numerous factors, including the concentration of clinicians in communities, nurse practitioner scope-of-practice laws, and patient characteristics such as insurance status. Our findings suggest that other group medical practices could incorporate more nonphysician staff without sacrificing productivity and thus profitability. However, the new staffing patterns that evolve may be affected by characteristics of the practice location or the types of patients served. ABSTRACT
N
umerous areas of the United States have an inadequate supply of primary care physicians, and the shortfall is expected to grow and spread.1 To address the shortage of primary care physicians and improve the quality of care delivered, health policy experts have recommended that medical practices transform themselves and embrace “team-based” care, encouraging a broader range of medical staff to work together.2–5 But to what extent should practices expand the use of nurse practitioners, physician assistants, nurses, or other medical staff, and what are the staffing implications? The paths to transforming primary care staffing and ambulatory care practices are not well mapped. Community health centers are federally
10.1377/hlthaff.2014.0098 HEALTH AFFAIRS 34, NO. 1 (2015): 95–103 ©2015 Project HOPE— The People-to-People Health Foundation, Inc.
doi:
Leighton Ku ([email protected]) is a professor in the Department of Health Policy, Milken School of Public Health, George Washington University (GWU), in Washington, D.C. Bianca K. Frogner is an associate professor in the Department of Family Medicine at the University of Washington, in Seattle. At the time of this research, she was an assistant professor in the Department of Health Services Management and Leadership, Milken School of Public Health, GWU. Erika Steinmetz is a senior research scientist in the Department of Health Policy, Milken School of Public Health, GWU. Patricia Pittman is an associate professor in the Department of Health Policy, Milken School of Public Health, GWU.
funded safety-net organizations that collectively provided comprehensive primary care to more than twenty-one million low-income patients in 2013.6 They have been in the vanguard of teambased care and flexibly vary the roles of medical staff based on local needs and staff capabilities. Community health centers use advanced-practice clinicians, such as nurse practitioners or physician assistants, more than regular private medical practices do.7 Finding and retaining physicians has long been a problem for community health centers, in part because they are located in medically underserved areas, which typically have shortages of primary care physicians.8 Community health centers have learned how to adjust their staffing while continuing to provide high-quality care in an affordable and efficient manner. CenJ a n u a ry 2 0 1 5
3 4: 1
Health Affai rs
95
Primary Care ter directors often say that they encourage staff to “work at the top of their capabilities,” taking greater responsibility for patient care consistent with their training, experience, and licensure, instead of working in rigidly defined roles. But staffing patterns can vary widely from one health center to another. An important factor in the extent to which staffing can be diversified is the effect of staffing patterns on productivity—a key measure of the efficiency of care. Higher productivity can reduce personnel costs or increase the number of visits, in either case increasing revenue from patients. Traditionally, productivity is measured based on the number of visits in which a physician (or advanced-practice staff member) sees a patient. This ignores the roles of other staff, however. A single clinician is not usually the sole provider of a patient’s care. A medical assistant may take vital signs; a physician may conduct the evaluation and make a diagnosis; and a nurse may draw blood, administer an injection, or educate the patient. According to a team-based care perspective, the visit should be viewed in terms of the joint productivity of the overall team. Instead of thinking of the visit as 100 percent produced by the physician, we might view 75 percent of the visit as produced by the physician, 10 percent by the medical assistant, and 15 percent by the nurse. Even staff who are not in the exam room play a role in medical productivity. For example, laboratory staff contribute to diagnoses, and quality assurance staff help monitor and improve care quality. Although they do not see patients, they are part of today’s medical practice system. This article examines different medical staffing patterns in community health centers, the impact of staffing on productivity, staff roles, and the factors that affect staffing patterns. The experiences of community health centers are relevant not only for safety-net providers, but also for the broader array of group medical practices in the United States as they implement team-based care approaches.
Study Data And Methods Data Our primary data source was the annual reports for 2012 filed with the Uniform Data System (UDS) by all community health centers that received federal section 330 grants, including sites in all fifty states, the District of Columbia, and US territories.6 Section 330 grants are the core federal grants provided to nonprofit community health centers by the Health Resources and Services Administration. All staff levels in these reports are expressed as full-time equivalents. The UDS also reports the 96
Health Affairs
J a n u a ry 2 0 1 5
3 4: 1
number of medical visits and patients (as opposed to dental or mental health visits and patients), diagnostic categories of visits, and other key caseload and financial characteristics of community health centers. There were 1,198 community health centers in the United States in 2012. We excluded seven whose reports were missing data, resulting in a final sample of 1,191 centers. Community health centers may operate clinics in multiple sites, and these 1,191 centers represented about 8,900 clinical sites in 2012. We classified medical staff into four categories: physicians (medical doctors and doctors of osteopathy; most were in primary care specialties), advanced-practice staff (nurse practitioners, physician assistants, and certified nurse-midwives), nurses (registered, licensed vocational, and practical nurses), and other medical staff (such as medical assistants, nurse aides, and lab or radiology staff). We excluded mental health, dental, vision, enabling service (for example, interpretation service), and administrative or finance visits and staff to focus on the core medical services available in most medical practices. We recognize the importance of mental health and dental care. However, most primary care practices do not employ behavioral health or dental staff.9 The UDS data also include the number of medical patients served at the health centers and the percentages of patients who are uninsured, are covered by Medicaid, and have limited English proficiency. We measured productivity as the number of weighted medical visits per center in 2012. Weighting is important because medical visits vary in the complexity of care required, which may also influence the type of staff involved. Diagnostic categories classified according to the International Classification of Diseases, Ninth Revision (ICD-9), codes were reported for about two-third of the UDS visits.We used data from the 2011 Medical Expenditure Panel Survey to compute average expenditures for office-based care for those diagnostic categories.We used the casemix of diagnoses and average expenditure per diagnosis to compute an expenditure index for each center. A center with a higher score tended to have more complex cases than a center with a low score (for more methodological information and analyses, see the online Appendix).10 By weighting visits, we sought to adjust for the complexity of different types of care required, depending on the diagnosis. To help understand workforce and policy environments that could affect staffing, we merged UDS data with county- and state-level data. We included county-level Area Health Resource File11
data about the concentrations of primary care physicians, advanced-practice staff, and nurses per 1,000 county residents and the percentage of the population in each county with incomes below the federal poverty level. We also included a rural-urban continuum code, ranging from 1 (the county was in a metropolitan area with a population of more than one million) to 9 (the most rural locale).12 We also merged the UDS data with state-level data on nurse practitioners’ scope of practice, or the range of services that nurse practitioners can conduct without a physician’s oversight in a given state. Nurse practitioners’ scopes of practice were categorized as full (nurse practitioners could treat patients or prescribe medications without a physician’s authorization), partial (nurse practitioners could treat but not prescribe), or restricted (most nurse practitioners’ actions must be authorized by a physician).13 Analysis To identify staffing patterns, cluster analysis was used. This is a method of pattern recognition that identifies clusters of similar observations—in this case, based on the percentages of staff in each of four categories in each health center. Each cluster was designed to differ maximally from the other clusters. We also examined the patterns of staffing by estimating correlation matrices. The UDS has limited information about who was present for each visit, showing only the type of lead clinician—usually a physician—who was responsible for the visit. To show the contributions of all medical staff and to compute joint productivity, we used the overall medical staff composition of the health center and number of weighted visits to compute the productivity of each staff type using multivariate ordinary least squares regression, with robust standard errors. This standard statistical approach, used in many related types of analyses, let us estimate the contributions of all medical staff, including some who might not see patients but who nonetheless contribute to the overall medical enterprise. Limitations This study was limited by the type and quality of data available. We could measure the number and types of staff at the community health centers. However, we could not measure their actual roles in patient care or how staff interacted to form effective teams. Further research is needed to obtain a more finely grained view of the roles and responsibilities of different types of health professionals working in teams. In addition, this study focused on the role of medical staff and did not include information about behavioral, dental, vision, enabling service, or administrative staff, who also have important roles to play. There may be some misclassification in staff
roles in the UDS. For example, a nurse employed at a community health center may serve in an administrative capacity or staff a health program that does not contribute to the reported number of medical visits (such as a Women, Infants, and Children nutrition program or a health education program). The UDS instructions ask that staff be classified by their main job function, but staff often serve multiple functions. A final limitation is that the approach we used to weight medical visits might not adequately reflect real differences in the amounts and types of resources across community health centers. However, comparisons with unweighted models suggested that weighting increased the effect of physicians, which is consistent with evidence that physicians tend to treat more complex cases than other staff do.
Study Results Staff Composition The median number of fulltime-equivalent medical staff at a community health center was 27.2, while the mean was 44.6. The median number of physicians at a health center was 5.0, and the mean was 8.8. This is comparable to the typical medical practice for American physicians: The median physician works in a group practice of 5–9 doctors.14 On average, physicians made up 19.0 percent of the medical staff at community health centers (Exhibits 1 and 2). But the size and staffing composition varied widely, as shown below. Patterns Of Staffing Using cluster analysis, we identified four dominant staffing patterns, Exhibit 1 Composition Of Medical Staff In Community Health Centers Overall And For Four Staffing Clusters, 2012
SOURCE Authors’ analysis of 2012 data from the Uniform Data System. NOTE Sample sizes are in Exhibit 2.
J a n u a ry 201 5
34:1
Health A ffairs
97
Primary Care Exhibit 2 Staffing And Other Characteristics, By Community Health Center (CHC) Staffing Clusters, 2012 Type of cluster Characteristic Number of CHCs
All CHCs 1,191
Typical 421
High advancedpractice staff 44
High nursing staff 295
High other medical staff 431
Mean percent of staff Physicians Advanced-practice Nurses Other medical
19.0% 17.9 25.6 37.4
19.1% 15.2 27.5 38.2
11.7% 54.4 19.2 14.7
21.2% 17.4 48.9 12.5
18.2% 17.2 8.6 56.1
44.6 2.58 15,137 50,234 1,123 52.1%
52.5 2.21 17,179 56,898 1,071 52.3%
10.4 0.24 5,161 13,545 1,261 38.6%
30.1 4.67 11,009 37,516 1,223 32.9%
50.1 1.85 16,985 56,192 1,091 66.6%
0.4% 38.7 34.5
15.0% 36.7 36.0
15.1% 46.9 28.6
14.2% 35.6 31.6
23.0% 41.9 35.4
Mean FTE medical staff Mean physicians per advanced-practice staff Medical patients Weighted medical visits Mean weighted visits per FTE medical staff Percent urban centers Mean percent of patients with: Limited English proficiency No insurance Medicaid
SOURCE Authors’ analysis of 2012 data from the Uniform Data System. NOTES All differences across the four groups are significant (p < 0:001). For details about the types of staff, see the “Data” section. FTE is full-time equivalent.
based on the percentages of medical staff in the four categories. The patterns were typical, high advanced-practice staff, high nursing staff, and high other medical staff. The typical group had a staff distribution similar to the overall national average (Exhibits 1 and 2). In the high advancedpractice staff group, about half of the staff were advanced-practice clinicians. Similarly, in the high nursing staff group, almost half of the staff were nurses, and in the high other medical staff group, about half of the staff were other medical professionals. The high advanced-practice staff group had a lower percentage of physicians than any other group. Across all centers, there were 2.58 physicians for every advanced-practice staff member (Exhibit 2). But in centers with high advancedpractice staff, the advanced-practice staff outnumbered physicians by about three to one. In high nursing staff centers, there were 4.67 physicians per advanced-practice staff member. The percentages of nurses and other medical staff at a community health center were strongly negatively correlated (r ¼ −0:84), meaning that centers with more nurses had fewer other medical staff. The percentages of physicians and advanced-practice staff were also negatively correlated (r ¼ −0:52), meaning that centers with more physicians had fewer advanced-practice staff (for the full correlation matrix, see the Appendix).10 The four clusters varied in a number of ways (Exhibit 2). Health centers in the typical and 98
H ea lt h A f fai r s
January 2015
34:1
high other medical staff clusters tended to be the largest, with roughly fifty medical staff, 17,000 medical patients, and 56,000 weighted medical visits, on average. High advanced-practice staff centers were the smallest. A basic measure of productivity is the number of weighted medical visits per staff member, without regard to staff type. The average levels were similar across the clusters, ranging from 1,071 to 1,261 visits per staff member (Exhibit 2). The levels were slightly higher for the high advanced-practice staff and high nursing staff centers. We did not have average salary information by category, so we could not determine the relative labor cost per visit. There may be a variation in the types of care provided in the different clusters, but it is important to remember that visits were weighted to adjust for differences in patients’ diagnoses. Of the four clusters, centers with high other medical staff were the most likely to be urban (66.6 percent) and to have the greatest share of patients with limited English proficiency (23.0 percent; Exhibit 2). Centers with high nursing staff were the most likely to be rural. Centers with high advanced-practice staff had the highest proportion of uninsured patients but the lowest percentage of patients with Medicaid. Contributions To Productivity Regression models were used to measure the contributions of different types of staff to overall productivity
(Exhibit 3). Not surprisingly, physicians consistently had the highest marginal contribution to productivity. Overall, one additional physician was associated with 2,994 additional visits. The marginal contribution of nurses appeared to be the lowest (292 additional visits), but the coefficient in that case was not significant. Because we were measuring joint productivity, the results were different than if we had used traditional measures of productivity. For example, there would have been 3,351 visits per physician if any physician-led visit was counted as belonging only to the physician, instead of the 2,994 visits estimated by our model, which assumed that there were also contributions by other staff. The marginal contributions model was also run separately for each of the four clusters. The results differed somewhat across cluster types. In high advanced-practice centers, advanced practitioners had higher contributions. Similarly, the effects of nurses and other medical staff were larger in high nurse or high other medical centers, respectively. For example, in the high nursing staff centers, an additional nurse was associated with 1,407 additional visits (Exhibit 3). In some models, the coefficients for nurses or for other medical staff were not significant. This might have occurred because the levels of nurses and of other medical staff were negatively correlated, so they canceled each other out. As a check, we developed models that pooled
nurses and other medical staff. In these consolidated models, the coefficients for the combined nurses and other medical staff were positive and significant overall and in all but one of the cluster-specific models. These results support the notion that these staff members contribute to productivity, but their negative correlation leads them to cancel each other out when both are in the models. As described in the Appendix,10 we tested alternative models using unweighted visits as the productivity metric. The results were generally similar. However, the weighted models’ coefficients for physicians tended to be higher, while those for advanced-practice staff were lower. This signals that weighting increases the apparent contribution of physicians and decreases the apparent contribution of advanced-practice staff, which suggests that physicians are more involved in complex visits, compared to the advanced-practice staff. Factors Affecting Staff Composition We also examined the factors that might explain why community health centers have varying staffing patterns. Exhibit 4 presents the results of regression models that examined the relationship of these factors to the percentages of medical staff who were physicians, advanced-practice staff, nurses, and other medical staff. We examined the roles of factors related to clinician availability, state policies, and health center caseloads. Rural or urban location was critical in all four models. The analyses show that rural location
Exhibit 3 Contributions To Medical Visit Productivity Per Staff Person, By Type Of Community Health Center (CHC) Staffing Cluster, 2012 Type of cluster
Physicians
Advancedpractice staff
Nursing staff
Other medical staff
Adjusted R2
All Coefficient 95% CI
2,994**** (1,955, 4,036)
1,584**** (1,394, 2,073)
292 (−119, 704)
548**** (352, 745)
0.952 —a
Typical Coefficient 3,370**** 95% CI (2,746, 3,995) High advanced-practice staff
1,546**** (907, 2,185)
347 (−241, 936)
265 (−115, 645)
0.963 —a
Coefficient 95% CI
2,761**** (1,877, 3,646)
2,287*** (974, 3,600)
4 (−1,159, 1,166)
−727 (−2,158, 703)
0.938 —a
2,086**** (1,317, 2,854)
198 (−734, 1,138)
1,407**** (922, 1,892)
357 (−239, 953)
0.948 —a
1,664**** (761, 2,567)
−788 (−1,837, 259)
744**** (322, 1,165)
0.945 —a
High nursing staff Coefficient 95% CI
High other medical staff Coefficient 2,923**** 95% CI (1,159, 4,686)
SOURCE Authors’ analysis of 2012 data from the Uniform Data System. NOTES The coefficients estimated by the models indicate the effect of adding one staff member on the number of weighted medical visits. For example, one additional advanced-practice staff member was associated with 1,584 additional visits. For details about the types of staff, see the “Data” section. CI is confidence interval. aNot applicable. ***p < 0:01 ****p < 0:001
January 2015
34:1
H e a lt h A f fai r s
99
Primary Care Exhibit 4 Coefficients For Factors Associated With Percentage Of Staff In Community Health Centers (CHCs), 2012 Dependent variable—percent of staff who are: Factor County characteristics
Physicians
Nursing staff
Other medical staff
−0.63****
0.76****
3.16**** −3.21**** 0.36*** 0.01
−3.46**** 4.86**** −0.54*** −0.14**
−2.09*** −0.38
2.02** 0.37
−7.80**** −0.35
7.87**** 0.35
0.01 −10.92**** −2.16 0.02
−0.10**** 10.96**** 0.88 −0.05***
0.01 −3.48 −6.44 −0.18****
0.08** 3.44 7.72 0.21****
24.30**** 0.109
15.95**** 0.151
25.49**** 0.185
34.26**** 0.204
Rural-urban continuum Per 1,000 population Physicians Advanced-practice staff Nurses Percent of population in poverty State scope-of-practice laws Full versus restricted scope Partial versus restricted scope CHC patients Number (1,000s) Percent uninsured Percent with Medicaid Percent with limited English proficiency
Advancedpractice staff
Constant Adjusted R2
2.13*** −0.08 1.31 0.43 −0.09
−2.26**** 0.38 −2.95** −0.24 0.22**
SOURCE Authors’ analysis of 2012 data from the Uniform Data System. NOTE For details about the types of staff, rural-urban continuum, and scope of practice laws, see the “Data” section. **p < 0:05 ***p < 0:01 ****p < 0:001
was associated with more advanced-practice staff and nurses, and with fewer physicians and other medical staff, being employed at a health center. Staffing was affected by the availability of health care professionals in the area. Community health centers in counties with more primary care physicians per 1,000 population had more physicians and fewer advanced-practice staff (Exhibit 3). Similarly, centers in counties with more advanced-practice staff per 1,000 population had more of that type of staff member and fewer physicians. Interestingly, centers in counties with higher concentrations of advancedpractice staff also had fewer other medical staff. A higher county-level concentration of nurses was associated with more physicians and fewer advanced-practice staff at centers but did not significantly affect the share of nurses on the staff. Nurse practitioner scope-of-practice laws were also associated with community health center staffing. We examined the effects of full and partial scopes of practice, compared to restricted scope (Exhibit 4). Centers in states with full scope of practice used slightly fewer physicians and slightly more advanced-practice staff. Full scope of practice was also associated with fewer non-advanced-practice nurses, but more other medical staff. Finally, we examined the characteristics of community health center caseloads (Exhibit 4). 100
Health Affairs
J a n u a ry 201 5
34:1
Other medical staff were more common in centers with more patients, while the opposite was true for advanced-practice staff. Community health centers with more uninsured patients had fewer physicians and more advanced-practice staff, but the percentage of patients with Medicaid was not significantly associated with staffing differences. The share of patients with limited English proficiency was associated with greater use of other medical staff.
Discussion Our analyses indicate that physicians are the leading contributors to productivity, but other medical staff also make meaningful contributions and increase the amount of care that can be delivered. Given concerns about primary care physician shortages, our findings suggest that greater use of nonphysician medical staff could enable more patients to be seen with limited resources. The average ratio of physicians to nonphysician medical staff for community health centers nationwide was 1:4. In comparison, a study of practices in a comprehensive primary care demonstration found a ratio of 1:2.5, which is still a higher ratio than that found in other practices.9 The ability to implement effective team-based primary care is relevant not only to the future of community health centers, but also to the broader system of group medical prac-
There is further room for innovation in the effective use of lowerlevel staff.
tice in the United States. We found diverse staffing patterns in community health centers, which suggests that staff roles and functions vary across settings. Across the four common staffing patterns that we identified, the overall levels of productivity per staff were relatively similar, even after we adjusted for the diagnostic complexity of the visits. There did not appear to be a distinctly optimal mix of staffing among the four patterns: Practices can be productive with varying blends of physicians, advanced-practice staff, nurses, and other medical staff. We found that local factors—such as rural location, the availability of clinicians in the area, scope-of-practice laws, and the patient and revenue mix of the centers—may be the most important determinants of staff composition. The data do not directly tell us what roles each type of staff play, but we can infer broad patterns. Nurses and other medical staff typically function as complements to physicians or advanced-practice staff, helping them reduce the time they spend on tasks that require less specialized skills. However, the strong negative correlation between nurses and other medical staff shows that more use of nurses is associated with less use of other medical staff, and vice versa. This result suggests that their roles may be somewhat interchangeable.6 In addition, nurses and other medical staff appear to make lower contributions to productivity than physicians or advanced-practice staff do. It is plausible that many of the functions that nurses and other medical staff fulfill—including patient education, quality assurance, and scheduling and administrative tasks—are necessary for care but not as strongly correlated with an increased number of medical visits, compared to the typical activities of physicians. Thus, nurses and other medical staff appear to have lower productivity. There is further room for innovation in the effective use of lower-level staff. For example, a recent study showed that the use of medical assistants as health coaches improved patient outcomes.15 More detailed research, including qualitative studies, could augment our under-
standing of the roles of nurses and other medical staff in primary care. We found a moderate negative correlation between physicians and advanced-practice clinicians, which also indicates some interchangeability of roles. In some cases, physicians and advanced-practice staff serve similar functions as primary care providers—for instance, both can make diagnoses or issue prescriptions— and thus can serve as substitutes for each other. Even in states with restrictive scope-of-practice laws, advanced-practice staff are able to diagnose and prescribe, with a physician’s authorization. In other cases, advanced-practice staff may serve as complements to physicians. For instance, a physician may see patients with more complex conditions or initiate care for a patient, while advanced-practice staff provide important follow-up care or patient education.2,9,16,17 A recent analysis of data from the National Ambulatory Medical Care Survey examined the use of physicians and advanced-practice staff in community health centers and found that advancedpractice staff were more involved than physicians with preventive patient education and mental health care visits.18 Our analyses support evidence that physicians tend to be more involved in care for patients with complex conditions, compared to nonphysician medical staff. Staffing patterns differ across community health centers, which indicates that staff functions also differ. But in general, regardless of the staffing mix, the centers’ overall productivity is relatively similar. Instead of having a one-sizefits-all approach to staffing and care provision, community health centers seem to be flexible and take advantage of the staff they have. Diverse staffing patterns are also related to geographic differences, such as rural or urban location, and the concentration of primary care physicians or advanced-practice staff in the area. Community health centers are located in areas that are designated by the Health Resources and Services Administration as Medically Underserved Areas, which typically have a low concentration of physicians. But physician workforce shortages affect other types of practices across the country, and the issue is likely to become even more pressing in the coming years. We also found other factors that affected staff distribution. Community health centers with more advanced-practice staff had more uninsured patients. A possible explanation is that centers with more uninsured patients are under financial strain that makes it harder for them to attract or retain physicians, and thus more advanced-practice staff are used. In addition, we found that community health centers with high levels of other medical staff had more patients January 2015
34:1
Health A ffairs
101
Primary Care with limited English proficiency, which suggests that these staff are providing additional services such as language interpretation. The distribution of staff differed in states where nurse practitioners had full scope of practice. Community health centers in these states had more advanced-practice staff and other medical staff, but fewer physicians and nurses. This is not necessarily a causal relationship: The adoption of scope-of-practice laws is not random, and it is possible that the relationship occurs because states with fewer physicians adopted full scope of practice to compensate. Regardless, it is reasonable to believe that granting nurse practitioners full scope of practice gives greater flexibility to community health centers and other medical practices in staffing choices, which is helpful if primary care physicians are in short supply. An Institute of Medicine report recommended expanding nurse practitioners’ scope of practice.19 We found that the use of advanced-practice staff was greatest in smaller community health centers that were more rural and had more uninsured patients. Previous studies have found that advanced-practice professionals can offer high-quality primary care services, often comparable to those provided by physicians.16,20,21 It is beyond the scope of this study to assess whether some staffing patterns are better or worse than others or lead to better patient outcomes or greater patient satisfaction. Other research has found that in general, community health centers provide high-quality primary care, broadly comparable to that offered in other ambulatory care settings, despite the resource limitations that health centers face and the low incomes and spotty insurance coverage of their patients.22–25 Community health centers have been in the vanguard of innovations in medical practice. As of 2013, 80 percent of community health centers had electronic health record systems, and 54 percent were recognized as patientcentered medical homes.26 The number of community health centers in the United States has grown rapidly for more than a decade and is expected to continue to grow. It seems likely that the centers will continue to innovate and be flexible in staffing and defining staff roles, while also providing highquality care.
Conclusion As the number of community health centers grows, leading to potential staffing pressures, centers must develop innovative staffing and
102
Health Affairs
J a n u a ry 201 5
34:1
The distribution of staff differed in states where nurse practitioners had full scope of practice.
care approaches to provide high-quality primary care in an efficient manner. The examples of community health centers may provide insights for other medical practices. In the broader US health care system, group medical practices are now facing many of the same pressures that community health centers have long encountered. There has been substantial research (and sometimes debate) about the broader roles of physicians and advanced-practice staff. However, there has been less attention to the important ways in which nurses and other medical staff contribute to care. Our findings indicate that it is possible to include a broad array of advanced-practice staff, nurses, and other medical staff in ambulatory medical practices to work alongside physicians and make meaningful contributions to productivity. Increasing the share of nonphysician staff may be more feasible for larger group practices than for small or solo practices because the larger size makes it possible to deploy staff in more diverse ways. Multiple staffing patterns, using more nonphysician staff in team-based care, are possible and could increase the number of patients served. In addition to regulatory issues such as scope of practice, insurers may need to consider whether their payment or credentialing policies are creating barriers to more flexible use of medical staff in patient care. For example, even if a state has full scope of practice for nurse practitioners, there may be problems if an insurer will pay only a physician for patient care services. We found that productivity levels were relatively similar across all staff configurations and that there did not appear to be a single optimal model of staffing. Both community health centers and the broader set of ambulatory practices can use diverse approaches that fit their needs and capacities and those of their communities. ▪
This research was presented at the AcademyHealth Annual Research Meeting in San Diego, California, June 10, 2014. This article was supported in part under a cooperative
agreement with the National Center for Health Workforce Analysis, Health Resources and Services Administration. The Uniform Data Systems data used in this article were provided by the Office
of Data and Quality, Bureau of Primary Health Care, Health Resources and Services Administration.
NOTES 1 Petterson SM, Liaw WR, Phillips RL, Rabin DL, Meyers DS, Bazemore AW. Projecting US primary care workforce needs: 2010–2025. Ann Fam Med. 2012:10(6):503–9. 2 Blumenthal D, Abrams MK. Putting aside preconceptions—time for dialogue among primary care clinicians. N Engl J Med. 2013;368(20):1933–4. 3 Chen EH, Bodenheimer T. Improving population health through teambased panel management: comment on “Electronic medical record reminders and panel management to improve primary care of elderly patients.” Arch Intern Med. 2011; 171(17):1558–9. 4 Bodenheimer TS, Smith MD. Primary care: proposed solutions to the physician shortage without training more physicians. Health Aff (Millwood). 2013;32(11):1881–6. 5 Auerbach DI, Chen PG, Friedberg MW, Reid R, Lau C, Buerhaus PI, et al. Nurse-managed health centers and patient-centered medical homes could mitigate expected primary care physician shortage. Health Aff (Millwood). 2013;32(11):1933–41. 6 Health Resources and Services Administration. Health center data [Internet]. Rockville (MD): HRSA; [cited 2014 Nov 17]. Available from: http://bphc.hrsa.gov/healthcenter datastatistics/ 7 Hing E, Hooker RS, Ashman JJ. Primary health care in community health centers and comparison with office-based practice. J Community Health. 2011;36(3):406–13. 8 Rosenblatt RA, Andrilla CH, Curtin T, Hart LG. Shortages of medical personnel at community health centers: implications for planned expansion. JAMA. 2006;295(9): 1042–9. 9 Peikes DN, Reid RJ, Day TJ, Cornwell DD, Dale SB, Baron RJ, et al. Staffing patterns of primary care practices in the Comprehensive Primary Care Initiative. Ann Fam Med. 2014;12(2):142–9. 10 To access the Appendix, click on the
11
12
13
14
15
16
17
18
Appendix link in the box to the right of the article online. Health Resources and Services Administration. Area Health Resources File [Internet]. Rockville (MD): HRSA; [cited 2014 Nov 24]. Available from: http://ahrf.hrsa.gov Economic Research Service. Rural/ urban continuum codes [Internet]. Washington (DC): Department of Agriculture; [last updated 2013 May 10; cited 2014 Nov 24]. Available from: http://www.ers.usda.gov/ data-products/rural-urbancontinuum-codes.aspx Pearson L. The 2012 Pearson report: a national overview of nurse practitioner legislation and health care issues. Monroe (NJ): NP Communications LLC; 2012. Kane CK, Emmons DW. New data on physician practice arrangements: private practice remains strong despite shifts in hospital employment [Internet]. Chicago (IL): American Medical Association; c2013 [cited 2014 Nov 12]. (Policy Research Perspectives). Available from: http:// www.nmms.org/sites/default/files/ images/2013_9_23_ama_survey_ prp-physician-practicearrangements.pdf Willard-Grace R. Not just a medical assistant anymore: new healthcare delivery models that use unlicensed professionals. Paper presented at: AcademyHealth Annual Research Meeting; 2014 Jun 9; San Diego, CA. Donelan K, DesRoches CM, Dittus RS, Buerhaus P. Perspectives of physicians and nurse practitioners on primary care practice. N Engl J Med. 2013;368(20):1898–906. Everett C, Thorpe C, Palta M, Carayon P, Bartels C, Smith MA. Physician assistants and nurse practitioners perform effective roles on teams caring for Medicare patients with diabetes. Health Aff (Millwood). 2013;32(11):1942–8. Downie S. Comparison of primary care visits to physicians and nonphysician providers in community
19
20
21
22
23
24
25
26
health centers and hospital outpatient departments. Paper presented at: AcademyHealth Annual Research Meeting; 2014 Jun 10; San Diego, CA. Institute of Medicine. The future of nursing: leading change, advancing health. Washington (DC): National Academies Press; 2011. Mundinger MO, Kane RL, Lenz ER, Totten AM, Tsai WY, Cleary PD, et al. Primary care outcomes in patients treated by nurse practitioners or physicians: a randomized trial. JAMA. 2000:283(1):59–68. Stanik-Hutt J, Newhouse RP, White KM, Johantgen M, Bass EB, Zangaro G, et al. The quality and effectiveness of care provided by nurse practitioners. J Nurse Pract. 2013;9(8): 492–500. Goldman LE, Chu PW, Tran H, Romano MJ, Stafford RS. Federally qualified health centers and private practice performance on ambulatory care measures. Am J Prev Med. 2012;43(2):142–9. Bruen BK, Ku L, Lu X, Shin P. No evidence that primary care physicians offer less care to Medicaid, community health center, or uninsured patients. Health Aff (Millwood). 2013;32(9):1624–30. Probst JC, Laditka JN, Laditka SB. Association between community health center and rural health clinic presence and county-level hospitalization rates for ambulatory care sensitive locations: an analysis across eight US states. BMC Health Serv Res. 2009;9:134. Richard P, Ku L, Dor A, Tan E, Shin P, Rosenbaum S. Cost savings associated with the use of community health centers. J Ambul Care Manage. 2012;35(1):50–9. Health Resources and Services Administration. 2013 health center data [Internet]. Rockville (MD): HRSA [cited 2014 Nov 12]. Available from: http://bphc.hrsa.gov/uds/ datacenter.aspx?year=2013
January 2015
34:1
Health A ffairs
103
