- Email: [email protected]
Advanced practice clinicians as a usual source of care for adults in the United States
Available online at www.sciencedirect.com
Nurs Outlook xxx (2016) 1e9
www.nursingoutlook.org
Advanced practice clinicians as a usual source of care for adults in the United States Matthew A. Davis, MPH, PhDa,*, Cui Guo, MSb, Marita G. Titler, PhD, RN, FAANa, Christopher R. Friese, PhD, RN, AOCN, FAANa a
Department of Systems, Populations, and Leadership, University of Michigan School of Nursing, Ann Arbor, MI Department of Biostatistics Graduate Program, University of Michigan School of Public Health, Ann Arbor, MI
b
article info
abstract
Article history: Received 19 February 2016 Revised 22 June 2016 Accepted 4 July 2016
Background: Advanced practice clinicians (APCs) including nurse practitioners and
Keywords: Nurse practitioners Physician assistants Primary health care Delivery of health care Health costs
physician assistants are increasingly used to deliver care, yet little is known about these providers as a usual source of primary care. Purpose: This study examined the extent to which APCs serve as a usual source of care and the impact of such use on health care expenditures and quality. Methods: We performed a cross-sectional study by identifying 90,279 adults from the 2002 to 2013 Medical Expenditure Panel Survey who self-reported their usual source of care as either an APC or a primary care physician (PCP). Using complex survey design methods to make national estimates, we compared annual health care expenditures and quality measures among adults whose usual source of care is an APC to that of adults whose usual source of care is a PCP. Discussion: Nationally, 32 million adults visit an APC each year, yet only 1.4 million adults report their usual source of care to be an APC. In adjusted analyses, mean annual health care expenditures were $7,323 among APC patients vs. $7,959 among PCP patients, a difference of $635 (95% confidence interval [$1,408 to $138]). Across specific health services, APC patients trended toward having lower expenditures except for marginally higher expenditures on emergency room visits ($256 vs. $227 p < .001). APC patients were similar to that of PCP patients across health care quality measures. Conclusions: Few U.S. adults report their usual source of care to be an APC. Health care spending and quality measures are similar between APC patients and PCP patients. Expanding use of APCs as a usual source of care will likely not lead to increased health care spending. Cite this article: Davis, M. A., Guo, C., Titler, M. G., & Friese, C. R. (2016, -). Advanced practice clinicians as a usual source of care for adults in the United States. Nursing Outlook, -(-), 1-9. http://dx.doi.org/ 10.1016/j.outlook.2016.07.006.
Introduction As health care costs continue to rise (Keehan et al., 2015) and a shortage of primary care physicians
(PCPs) is evident (Bodenheimer & Pham, 2010; Ku, Jones, Shin, Bruen, & Hayes, 2011; West & Dupras, 2012), policy makers seek new strategies to meet the growing health care needs of the nation. The Patient Protection and Affordable Care Act contains several
* Corresponding author: Matthew A. Davis, University of Michigan, 400 North Ingalls, Room 4347, Ann Arbor, MI 48109. E-mail address: [email protected] (M.A. Davis). 0029-6554/$ - see front matter Ó 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.outlook.2016.07.006
2
Nurs Outlook xxx (2016) 1e9
provisions to improve access to care and expand health insurance coverage to millions of previously uninsured Americans. Upward of 7,200 additional physicians are required to provide primary care to the Americans who obtain coverage through the Affordable Care Act alone (Hofer, Abraham, & Moscovice, 2011; Huang & Finegold, 2013). Accounting for the aging population coupled with expansion of health insurance coverage, the U.S. Department of Health and Human Services projects a total shortfall of 20,400 PCPs by 2020 (US Department of Health and Human Services, 2013). Among others, the National Academy of Medicine advocates for increasing the use of advanced practice clinicians (APCs)dnurse practitioners and physician assistants to address the current and projected demand for primary care services (Health Affairs Health Policy Brief, 2012; Institute of Medicine, 2010; Kaiser Family Foundation, 2015; Naylor & Kurtzman, 2010; US Department of Health and Human Services, 2013). APCs currently represent a quarter of the U.S. primary care workforce and practice in diverse settings, which include physician-led health care teams, retail clinics, and as independent providers (Auerbach, 2012; Hooker, Brock, & Cook, 2016; Kralewski, Dowd, Curoe, Savage, & Tong, 2015; Roblin, Howard, Becker, Kathleen Adams, & Roberts, 2004; Stange, 2014; Thygeson, Van Vorst, Maciosek, & Solberg, 2008). Although they perform similar primary care services, insurers generally reimburse APCs at lower rates than physicians (Health Affairs Health Policy Brief, 2012; Naylor & Kurtzman, 2010; Stange, 2014), which may result in APC-delivered primary care costing less than that of PCP. However, lower direct costs associated with APC care may be offset by additional indirect costs because of greater reliance on diagnostic services (Hughes, Jiang, & Duszak, 2015), consultations to specialty physicians, and emergency department visits. In 2010, the National Academy of Medicine recommended that nurse practitioners (the largest constitute of the APC workforce) practice to the full extent of their training as independent clinicians (Institute of Medicine, 2010). However, to date, very little is known about the extent to which APCs practice as a usual source of care and the question remains whether APC-directed primary care differs in meaningful ways compared to PCP-directed care in regards to spending and quality. Therefore, we used nationally representative data from the Medical Expenditure Panel Survey (MEPS) to examine the use of APCs as a usual source of care in the United States. Our objectives were to (a) estimate the number of American adults who visit APCs and those who specifically self-report their usual source of care to be an APC, (b) compare the characteristics of U.S. adults whose usual source of care is an APC to that of adults whose usual source of care is a PCP, and (c) compare annual health care expenditures and quality of care among adults whose usual source of care is an APC to that of adults whose usual source of care is a PCP.
Methods Study Population and Data Source For this study, we pooled 2002 to 2013 data from the MEPS. Conducted annually by the Agency for Healthcare Research and Quality, the MEPS is a nationally representative sample of the noninstitutionalized U.S. population (Cohen, Cohen, & Banthin, 2009). Detailed information is gathered on health care utilization, expenditures, and health status. The MEPS utilizes an overlapping panel design consisting of household, medical provider, and insurance provider components. For each year, personal- and family-level data obtained from the household, medical provider, and insurance provider are collected and aggregated by the MEPS study team. For this study, we used data from the MEPS household component files including the full-year consolidated data and medical conditions files. This study was granted an exemption from institutional board review. We analyzed data from all adult (18 years and older) participants of the MEPS survey from 2002 to 2013. For the study time period, a total of 295,701 adults participated, with a range of 21,782 adults in 2007 to a high of 27,820 in 2012. From 2002 to 2013, the response rates among all participants ranged between 58.6% and 69.3%.
Measures Adults Who Visit APCs To determine how common APC care for adults is in the United States, we identified annual ambulatory visits to APCs (any visit to either a nurse practitioner or physician assistant in the calendar year) using data from the MEPS consolidated data files. From this, we estimated the total number of U.S. adults who use APC care each year.
Identification of Usual Source of Care Next, we used MEPS participants’ self-reported usual source of care as a surrogate measure of their primary care provider type (Friedberg, Hussey, & Schneider, 2010). MEPS participants are asked whether there is a particular doctor’s office, clinical, health center, or other place the individual usually goes to when sick or in need of health advice. From those who answered “yes,” we estimated the number of U.S. adults who have a usual source of care. For those who have a self-reported usual source of care, MEPS staff then inquires about whether their usual source of care is a “facility,” “person,” or “person within facility.” Those who report having a specific provider (i.e., either a person or person within a health care facility) are then asked to identify the professional credentials of the provider. We identified MEPS participants whose usual source of care was an APC
Nurs Outlook xxx (2016) 1e9
(defined as a “nurse practitioner,” or “physician assistant”) vs. PCP (defined as “MD, general/family practice” or “MD, internal medicine”). Our final analytic sample consisted of 90,279 adults who reported using either an APC or PCP as a usual source of care. Herein, “APC patients” refers to adults whose usual source of care is either a nurse practitioner or physician assistant, whereas “PCP patients” refers to adults whose usual source of care is a PCP.
Annual Health Care Expenditures In 6-month follow-up increments, MEPS participants are asked about all of their health care use such as ambulatory visits, inpatient and outpatient care, and prescription medications. MEPS staff then gathers information on the details regarding the health care use from providers and records how much was spent on the service. Aggregating across all health services, MEPS estimates the total amount spent in the year on health care servicesdwe used this as our primary end point. We extracted annual expenditure data on specific health services including office based, outpatient, prescription medications, emergency department visits, and inpatient care.
Health Care Quality Measures In addition to collecting information about health care expenditures, MEPS also captures a variety of selfreported health care quality measures. We adapted four primary care quality measures from the U.S. Health and Human Services core set of health care quality measures including the percentage of (a) adults between the ages 18 to 64 years who received an annual seasonal flu vaccination, (b) smokers who were encouraged by their provider to stop, (c) women between the ages 21 and 64 years who received a Pap smear examination within the past 3 years, and (d) diabetic patients who had an annual hemoglobin A1c (HbA1c) laboratory test (Kim, Park, Cha, & Jeong, 2015).
Characteristics of Study Participants We extracted sociodemographic data for participants in our study including age, sex, race/ethnicity (nonHispanic white, non-Hispanic black, Hispanic, vs. other), marital status (married, widowed/divorced/ separated, vs. never married), level of education (high school or less vs. above high school), U.S. region of residence (Northeast, Midwest, South, vs. West), health insurance coverage (private, public, vs. uninsured), and rurality. Metropolitan and micropolitan statistical areas (MSA; US Census Bureau, 2013) were used as a measure of ruralitydparticipants who resided in an MSA were considered nonrural vs. those residing in non-MSA areas were deemed to be in rural locales. As MEPS participants are surveyed multiple times over the calendar year, for measures that were acquired multiple times per year, we used the last measurement of the corresponding calendar year. As it is a strong predictor of health and mortality (DeSalvo, Bloser, Reynolds, He, & Muntner, 2006), we
3
used self-reported health status as an overall measure of health and collapsed it into “excellent, very good, or good” vs. “fair or poor.” Based on a combination of both physical and cognitive limitations measures, we also calculated the percentage of participants with “any functional limitation.” The MEPS also collects information on body mass index (BMI), smoking status, and administers the 12-item Short Form Survey (SF-12) to participants. We determined the percentage of participants who were obese (BMI 30 kg/m2) and the percentage who smoked, and we calculated the mean physical and mental summary component scores (PCS and MCS) that range from 0 to 100 (100 indicating the highest level of health). Finally, for each participant, based on all medical conditions coded by the International Classification of Diseases, version 9, Clinical Modification, we generated a comorbidity score using a modified version of the Charlson Comorbidity Index (Charlson, Pompei, Ales, & MacKenzie, 1987; D’Hoore, Bouckaert, & Tilquin, 1996).
Statistical Analyses Our primary analyses examined the relationship between usual source of care type (APC vs. PCP) and annual health care expenditures. Some study participants had no spending, and data from those with health care expenditures were highly skewed. Therefore, we used established modeling methods for zero inflated, skewed data that included a two-part multivariate model to generate predicted annual expenditures (Phillips et al., 2009). First, we used a logistic model to predict the probability of any spending. Second, among participants who had expenditures, we used a generalized linear model with log-link function and gamma distribution to predict annual expenditures. The final estimate of predicted annual spending was calculated by multiplying the predictions from the two models together. We did not use smearing methods because the residuals were heteroskedastic (Duan, 1983). For all analyses, we used complex survey design methods to make national estimates, which account for a participant’s probability of selection and sampling methodology. To increase the sample size of APC patients for analysis, we aggregated data across the 13 years and adjusted survey weights accordingly so that our final analytic data set represented 1 calendar year. To account for inflation, we used the Consumer Price Index from the U.S. Bureau of Labor Statistics for medical services to convert all expenditure data to 2013 dollars (Cavazos-Rehg et al., 2015). Our final models included covariates for sociodemographic differences, number of chronic conditions (based on the Charlson index), U.S. region (a surrogate measure of scope of APC practice [Health Affairs Health Policy Brief, 2012; Spetz, Parente, Town, & Bazarko, 2013]), and calendar year (to adjust for difference in practice over time). We used a t test to compare means and a c2 test to compare proportions between APC and PCP patient. Analyses were based on complete case analysis, and we
4
Nurs Outlook xxx (2016) 1e9
assumed any missing values to be missing completely at random. A two-sided p value of <.05 was considered statistically significant. All analyses were conducted using Stata MP, version 14.0 (College Station, TX).
Table 1 e Characteristics of U.S. Adults by Usual Source of Care Type Characteristic Usual Source p of Care Type Value* APC
Findings Nationally, we estimate approximately 32 million adults (14.2% of the 226 million U.S. adults) visit an APC each year. However, only 1.4 million adults (<1% among the 171 million adults who have a usual source of care) use APCs as a usual source of care, whereas 79.4 million use PCPs (Table 1).
Characteristics of APC vs. PCP Patients The characteristics of adults who report their usual source of care to be an APC differed in several ways from adults whose usual source of care is PCPs. Adult APC patients were younger, more likely to be female, and more likely to be non-Hispanic white than PCP patients ( p < .001 for all comparisons; Table 1). The percentage of adults residing in rural areas was more than doubled for APC patients compared to PCP patients ( p < .001). Finally, APC patients were more likely to be uninsured than PCP patients ( p < .001). APC patients were more likely to have poor or fair self-reported health status, to be obese, and to smoke than PCP patients (Figure 1). The mean PCS of the SF-12 was 47.4 (standard error [SE]: 0.52) among APC patients compared to 48.1 (SE: 0.08) among PCP patients ( p ¼ .20), whereas the mean MCS of the SF-12 was 49.8 (SE: 0.34) for APC patients compared to 51.2 (SE: 0.06) for PCP patients ( p < .001).
Annual Health Care Expenditures by Usual Source of Care Type In unadjusted analyses, the mean annual health care expenditures were $6,760 among patients whose usual source of care was an APC vs. $6,719 among patients whose usual source of care was a PCP, a difference of $41 (95% confidence interval [CI] [909 to 991]), p ¼ .93 (Table 2). Across specific expenditures, APC patients appeared to have equivalent or higher health care expenditures compared to PCP patients, albeit the only statistically significant difference was in annual expenditures on prescription medicationdAPC patients spent $337 (95% CI [58 to 615]), p ¼ .02, more per year on prescription medications than PCP patients. In adjusted models, mean annual health care expenditures were $7,323 among patients whose usual source of care was APC compared to $7,959 among patients whose usual source of care was a PCP, a difference of $635 (95% CI [1,408 to 138]), p ¼ .11. The only statistically significant differences between APC and PCP patients were expenditures on outpatient
Sample size, number 1,444 National estimates, millions (%)y Number of adults 1.4 Sociodemographic characteristics Age category (years) 18 to <45 0.6 (41.8) 45 to <65 0.6 (40.2) 65 0.3 (18.0) Sex Male 0.4 (31.6) Female 1.0 (68.4) Race Non-Hispanic white 1.2 (86.4) Non-Hispanic black 0.1 (5.6) Hispanic 0.1 (4.4) Other 0.0 (3.5) Marital status Married 0.8 (55.5) Widowed/divorced/ 0.3 (24.5) separated Never married 0.3 (20.0) Education High school or less 0.7 (48.9) education Above high school 0.7 (51.1) education U.S. census region Northeast 0.3 (19.7) Midwest 0.3 (20.6) South 0.5 (35.6) West 0.3 (24.1) MSA, rurality Non-MSA 0.4 (37.1) MSA 0.8 (62.9) Health care insurance Any private 1.0 (70.2) Public only 0.3 (20.5) Uninsured 0.1 (9.3)
PCP 88,835 79.4
<.001 29.8 (37.5) 30.5 (38.4) 19.1 (24.0) <.001 36.0 (45.4) 43.3 (54.6) <.001 59.8 (75.4) 7.8 (9.7) 6.9 (8.7) 4.9 (6.2) .14 47.0 (59.2) 17.7 (22.2) 14.7 (18.5) .11 35.6 (44.9) 43.7 (55.1) .17 19.3 (24.2) 17.3 (21.8) 29.0 (36.6) 13.8 (17.4) <.001 12.3 (15.6) 66.6 (84.4) <.001 61.1 (76.9) 13.7 (17.2) 4.6 (5.8)
Note. APC, advanced practice clinicians; MSA, metropolitan statistical area; PCP, primary care physician. * Difference compares APC vs. PCP patients. c2 test used to compare proportions. y Estimates based on weighted sample using complex design methods.
services and emergency department visits. Annually, APC patients spent $51 (95% CI [95 to 6]) less on outpatient services and $28 (95% CI [14 to 43]) more on emergency room visits than PCP patients.
Quality of Care Across the four measures of adult primary care quality, APC patients were comparable to PCP patients (Figure 2). The only marginal differences were a higher percentage of APC patients who reported receiving seasonal flu vaccination and diabetics who received an annual HbA1c compared to PCP patients ( p ¼ .10 and .07 respectively).
5
Nurs Outlook xxx (2016) 1e9
50 APC 45
PCP
< 0.001
40 35 30
0.28
Percent 25
< 0.01 0.04
20 15
0.46
10 5 0 1 Functional Poor or fair limitations health status
Obese
Smoker
2 Chronic conditions
Figure 1 e Health status of U.S. adults by a usual source of care type. APC, advanced practice clinician; PCP, primary care physician.
Discussion To our knowledge, this is the first national study to examine APCs as a usual source of adult care as compared to PCPs. We found that although 32 million adults visit APCs at least once during the year (14.2%),
100 90
only 1.4 million (<1%) report their usual source of care to be an APC. Despite the sizeable contribution of APCs to the primary care workforce and the 2010 National Academy of Medicine’s recommendation that nurse practitioners (the predominate APC provider) practice to their full independent potential (Institute of Medicine, 2010), we were surprised to find so few U.S.
0.07
APC PCP
0.29
80 0.37
70 60
Percent 50 0.10
40 30 20 10 0 Received seasonal flu immunization a
Smokers Women who had Pap Diabetics who had encouraged to stop smear within 3 years b annual Hb A1c c
Figure 2 e Patient-reported adult primary care quality measures by usual source of care type. aRestricted to adults aged between 18 and 64 years. bRestricted to women aged between 21 and 64 years. cRestricted to diabetics between 18 and 65 years. APC, advanced practice clinician; HbA1c, hemoglobin A1c; PCP, primary care physician.
6
Nurs Outlook xxx (2016) 1e9
Table 2 e Difference in Mean Annual Healthcare Expenditures by Usual Source of Care Type* Health Service
Unadjusted All health services Office-based care Outpatient services Prescription medications Emergency department visits Inpatient services Adjustedx All health services Office-based care Outpatient services Prescription medications Emergency department visits Inpatient services
Mean Expenditure by Usual Source of Care Type
Difference (95% CI)y
p Valuez
APC
PCP
$6,760 $1,598 $662 $1,929 $232 $1,938
$6,719 $1,564 $660 $1,592 $206 $1,870
$41 (909 to 991) $34 (192 to 261) $34 (192 to 260) $337 (58 to 615) $26 (11 to 62) $68 (765 to 901)
.93 .77 .97 .02 .17 .87
$7,323 $1,713 $645 $1,826 $256 $1,870
$7,959 $1,739 $695 $1,965 $227 $1,978
$635 (1,408 to 138) $26 (144 to 92) $51 (95 to 6) $139 (375 to 97) $28 (14 to 43) $108 (286 to 70)
.11 .67 .03 .25 <.001 .23
Note. 95% CI, 95% confidence interval; APC, advanced practice clinicians; PCP, primary care physician. * All estimates based on weighted sample using complex design methods. y Difference compares mean annual expenditures between APC vs. PCP patients (i.e., mean expenditurephysician mean expendituresAPC). z T test used to compare means. x Adjusted for age, sex, race/ethnicity, obesity status, smoking status, calendar year, U.S. region, and number of chronic conditions (Charlson index score).
adults reporting their usual source of care as an APC. Previous studies have found that access to primary care services is insensitive to variation in APC availability and more favorable nurse practitioner scope of practice (Cross & Kelly, 2015; Stange, 2014). This suggests that APCs primarily function as participants in collaborative primary care delivery models, rather than as independent providers who serve as an alternative portal of entry to the health care system. Another possibility is that Americans’ views regarding the use of APCs have not yet shifted to reflect their contribution as independent primary care providers, thus underestimating APCs’ potential contributions. As the scope of APC practice legislation expands across states, it may be worth examining public perceptions of APCs as usual sources of care. Our findings that patients whose usual source of care was an APC are more likely to be female, younger, uninsured, and of poorer health status are congruent with prior research (DesRoches et al., 2013). Similar to prior studies, we found that adults who use APCs were more likely to reside in rural areas suggesting APCs care for traditionally underserved populations (Everett, Schumacher, Wright, & Smith, 2009; Grumbach, Hart, Mertz, Coffman, & Palazzo, 2003; Larson, Palazzo, Berkowitz, Pirani, & Hart, 2003). Collectively, these findings support the National Academy of Medicine’s report that APCs care for vulnerable populations in rural areas where primary care services by PCPs are limited. Variation in access to primary care already exists across the country and 7 million Americans live in places where the demand for primary care services greatly exceeds supply (Huang & Finegold, 2013). Projected growth in the number of APCs may help offset
the demand for PCP services in these locales; however, variation exists in the supply of APCs as well (Kaiser Family Foundation, 2015; US Department of Health and Human Services, 2013). Increasing the availability of APCs through innovative educational programs is needed to meet these identified demands (Institute of Medicine, 2010). Patients whose usual source of care was an APC had comparable spending (if not lower) to those with a PCP as a usual source of caredthe exception being marginally lower annual outpatient expenditures and slightly higher expenditures on emergency room visits. A previous report found that health care spending varies by usual source of care with higher spending among adults who seek care from medical specialists (Phillips et al., 2009). Although not statistically significant ( p ¼ .11), in adjusted analyses, annual health care expenditures among adults whose usual source of care was an APC were $635 lower than those who see PCPs. This finding parallels a recent report of older adults that found those who sought most care from nurse practitioners had 29% less spending on evaluation and management compared to those who see medical physicians (Perloff, DesRoches, & Buerhaus, 2016). Such differences in spending may suggest large effects on national health care costs. Based on our estimate of the difference in annual spending among the 1.4 million adults who use APCs as a usual source of care, the net difference on national health care spending would be a reduction in approximately $890 million per year, however, it could range from a decrease of $2.0 billion to an increase in $193 million. APCs receive lower reimbursement than physicians for similar services (Health Affairs Health Policy Brief,
Nurs Outlook xxx (2016) 1e9
2012; Naylor & Kurtzman, 2010; Stange, 2014). Although not equitable, lower reimbursement among APCs may be the mechanism underpinning lower costs observed among APC-delivered care compared to that of PCPs (Martin-Misener et al., 2015). Prior studies do not account for potential indirect effects such as reliance on other services when compare APC to PCP care (Hughes et al., 2015). One advantage of our approach is that we were able to capture the expenditures on specific health services that suggest lower spending on prescription medications and outpatient services (many of which are diagnostic procedures such as imaging) but slightly higher spending on emergency department services. As these analyses were adjusted for patient characteristics, this suggests some differences in clinical practice styles between APCs and PCPs. Our findings that APC care is equivalent to that of physicians echo a large literature dating back to the 1980s (Naylor & Kurtzman, 2010; Newhouse et al., 2011; Swan, Ferguson, Chang, Larson, & Smaldone, 2015). Previous studies have investigated the management of specific diseases by APCs including diabetes and encouraging healthy behaviors such as smoking cessation (Dierick-van Daele, Metsemakers, Derckx, Spreeuwenberg, & Vrijhoef, 2009; Hughes et al., 2015; Kuo, Chen, Baillargeon, Raji, & Goodwin, 2015; Mundinger et al., 2000; Virani et al., 2015). Our study contributes to this body of work by finding that APC care appears to be comparable to that of PCPs in regard to current aggregate measures of health care quality based on U.S. estimates (Kim et al., 2015; Newhouse et al., 2011).
Limitations Our study has several limitations that must be acknowledged. First, the premise of our study was based on the assumption that the usual source of care oversees much of the care patients receive (Phillips et al., 2009). Therefore, our analyses did not directly account for the provider who ordered diagnostic tests or directed medical referrals. The potential mixing of provider types participating in care delivery would likely bias our results toward the null, hence making the estimates reported here conservative. Second, MEPS data are self-reported by participants, and therefore inaccuracies could have introduced measurement error. Nevertheless, in the assignment of patients to provider types and on quality measures, self-report is advantageous over other methods of assignment that rely on assumptions made by investigators. Third, the study population was limited to noninstitutionalized U.S. adult citizens; thus, findings among children or institutionalized adults may differ. Many APCs provide pediatric care, and our analyses did not examine associations among children. Fourth, because of the sample size of MEPS respondents who reported their usual source of care to be either a nurse practitioner or physician assistant, our definition of APC combined nurse practitioners and physician assistants. While consistent with recent approaches
7
(Hughes et al., 2015), our analyses are therefore not able to identify potential differences between nurse practitioners and physicians assistants. Finally, as our study is an observational cross-sectional design, we cannot rule out the potential impacts of residual confounding on our findings.
Future Recommendations Findings from this study and those of other investigators noted above provide a scientific basis for a call to action for state policy legislation that promotes APCs to practice to the full scope for which they are prepared. Second, findings support those of other investigators that APCs quality of care is equivalent to that delivered by PCPs. Given this equivalency in quality of care, it is concerning that the number of U.S. adults reporting APCs as the usual source of care is so low and calls for studies to explicate perceptions of U.S. adults regarding services and quality of care delivered by APCs in primary care. In addition, it is recommended that collaborations are fostered with selected patient advocacy groups to increase their understanding and value of services provided by these clinicians. The sociodemographic characteristics of patients cared for by APCs (rural, uninsured, and those with poorer health status, particularly mental health) suggest that these providers are filling a gap in health care, and at a lower or comparable cost to PCPs. In particular, there are innovative models of care delivery for adults with chronic conditions that include APCs in development. For instance, in Australia, for patients with chronic illness, nurse practitioners routinely play a key role in ongoing management (Watts et al., 2009). The findings from our study call for innovative and accelerated education of APCs who can meet the demands for primary care and whose educational trajectory is shorter than those for PCPs. To fill the projected gap in primary care of the future, federal and state policies must address the need for APCs to practice to their full potential and increase the funding of education for APCs.
references
Auerbach, D. I. (2012). Will the NP workforce grow in the future? New forecasts and implications for healthcare delivery. Medical Care, 50(7), 606e610. Bodenheimer, T., & Pham, H. H. (2010). Primary care: Current problems and proposed solutions. Health Affairs (Millwood), 29(5), 799e805. Cavazos-Rehg, P. A., Krauss, M., Fisher, S. L., Salyer, P., Grucza, R. A., & Bierut, L. J. (2015). Twitter chatter about marijuana. Journal of Adolescent Health, 56(2), 139e145. Charlson, M. E., Pompei, P., Ales, K. L., & MacKenzie, C. R. (1987). A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. Journal of Chronic Diseases, 40(5), 373e383. Cohen, J. W., Cohen, S. B., & Banthin, J. S. (2009). The Medical Expenditure Panel Survey: A national information resource to
8
Nurs Outlook xxx (2016) 1e9
support healthcare cost research and inform policy and practice. Medical Care, 47(7 Suppl 1), S44eS50. Cross, S., & Kelly, P. (2015). Access to care based on state nurse practitioner practice regulation: Secondary data analysis results in the Medicare population. Journal of the American Association of Nurse Practitioners, 27(1), 21e30. D’Hoore, W., Bouckaert, A., & Tilquin, C. (1996). Practical considerations on the use of the Charlson Comorbidity Index with administrative data bases. Journal of Clinical Epidemiology, 49(12), 1429e1433. DeSalvo, K. B., Bloser, N., Reynolds, K., He, J., & Muntner, P. (2006). Mortality prediction with a single general self-rated health question. A meta-analysis. Journal of General Internal Medicine, 21(3), 267e275. DesRoches, C. M., Gaudet, J., Perloff, J., Donelan, K., Iezzoni, L. I., & Buerhaus, P. (2013). Using Medicare data to assess nurse practitioner-provided care. Nursing Outlook, 61(6), 400e407. Dierick-van Daele, A. T., Metsemakers, J. F., Derckx, E. W., Spreeuwenberg, C., & Vrijhoef, H. J. (2009). Nurse practitioners substituting for general practitioners: Randomized controlled trial. Journal of Advanced Nursing, 65(2), 391e401. Duan, N. (1983). Smearing estimate: A nonparametric retransformation method. Journal of the American Statistical Association, 78(383), 605e610. Everett, C. M., Schumacher, J. R., Wright, A., & Smith, M. A. (2009). Physician assistants and nurse practitioners as a usual source of care. Journal of Rural Health, 25(4), 407e414. Friedberg, M. W., Hussey, P. S., & Schneider, E. C. (2010). Primary care: A critical review of the evidence on quality and costs of health care. Health Affairs (Millwood), 29(5), 766e772. Grumbach, K., Hart, L. G., Mertz, E., Coffman, J., & Palazzo, L. (2003). Who is caring for the underserved? A comparison of primary care physicians and nonphysician clinicians in California and Washington. Annals of Family Medicine, 1(2), 97e104. Health Policy Brief. (2012). Nurse practitioners and primary care. Health Affairs (Millwood). Retrieved from http://www. healthaffairs.org/healthpolicybriefs/brief.php?brief_id=79. Hofer, A. N., Abraham, J. M., & Moscovice, I. (2011). Expansion of coverage under the Patient Protection and Affordable Care Act and primary care utilization. Milbank Quarterly, 89(1), 69e89. Hooker, R. S., Brock, D. M., & Cook, M. L. (2016). Characteristics of nurse practitioners and physician assistants in the United States. Journal of the American Association of Nurse Practitioners, 28(1), 39e46. Huang, E. S., & Finegold, K. (2013). Seven million Americans live in areas where demand for primary care may exceed supply by more than 10 percent. Health Affairs (Millwood), 32(3), 614e621. Hughes, D. R., Jiang, M., & Duszak, R., Jr. (2015). A comparison of diagnostic imaging ordering patterns between advanced practice clinicians and primary care physicians following office-based evaluation and management visits. JAMA Internal Medicine, 175(1), 101e107. Institute of Medicine. (2010). The Future of Nursing: Leading Change, Advancing Health. Washington, DC: The National Academies Press. Kaiser Family Foundation. (2015). Issue brief: Tapping nurse practitioners to meet the rising demand for primary care. Retrieved from http://kff.org/medicaid/issue-brief/tapping-nursepractitioners-to-meet-rising-demand-for-primary-care/. Keehan, S. P., Cuckler, G. A., Sisko, A. M., Madison, A. J., Smith, S. D., Stone, D. A., & Lizonitz, J. M. (2015). National health expenditure projections, 2014-24: Spending growth faster than recent trends. Health Affairs (Millwood), 34(8), 1407e1417. Kim, H., Park, J., Cha, M., & Jeong, J. (2015). The effect of bad news and CEO apology of corporate on user responses in social media. PLoS One, 10(5), e0126358.
Kralewski, J., Dowd, B., Curoe, A., Savage, M., & Tong, J. (2015). The role of nurse practitioners in primary healthcare. American Journal of Managed Care, 21(6), e366ee371. Ku, L., Jones, K., Shin, P., Bruen, B., & Hayes, K. (2011). The states’ next challengeesecuring primary care for expanded Medicaid populations. New England Journal of Medicine, 364(6), 493e495. Kuo, Y. F., Chen, N. W., Baillargeon, J., Raji, M. A., & Goodwin, J. S. (2015). Potentially preventable hospitalizations in Medicare patients with diabetes: A comparison of primary care provided by nurse practitioners versus physicians. Medical Care, 53(9), 776e783. Larson, E. H., Palazzo, L., Berkowitz, B., Pirani, M. J., & Hart, L. G. (2003). The contribution of nurse practitioners and physician assistants to generalist care in Washington State. Health Services Research, 38(4), 1033e1050. Martin-Misener, R., Harbman, P., Donald, F., Reid, K., Kilpatrick, K. , Carter, N., . DiCenso, A. (2015). Cost-effectiveness of nurse practitioners in primary and specialised ambulatory care: Systematic review. BMJ Open, 5(6), e007167. Mundinger, M. O., Kane, R. L., Lenz, E. R., Totten, A. M., Tsai, W. Y., Cleary, P. D., . Shelanski, M. L. (2000). Primary care outcomes in patients treated by nurse practitioners or physicians: A randomized trial. JAMA, 283(1), 59e68. Naylor, M. D., & Kurtzman, E. T. (2010). The role of nurse practitioners in reinventing primary care. Health Affairs (Millwood), 29(5), 893e899. Newhouse, R. P., Stanik-Hutt, J., White, K. M., Johantgen, M., Bass, E. B., Zangaro, G., . Weiner, J. P. (2011). Advanced practice nurse outcomes 1990-2008: A systematic review. Nursing Economics, 29(5), 230e250. Perloff, J., DesRoches, C. M., & Buerhaus, P. (2016). Comparing the cost of care provided to Medicare beneficiaries assigned to primary care nurse practitioners and physicians. Health Services Research, 51(4), 1407e1423. Phillips, R. L., Dodoo, M. S., Green, L. A., Fryer, G. E., Bazemore, A. W., McCoy, K. I., & Petterson, S. M. (2009). Usual source of care: An important source of variation in health care spending. Health Affairs (Millwood), 28(2), 567e577. Roblin, D. W., Howard, D. H., Becker, E. R., Kathleen Adams, E., & Roberts, M. H. (2004). Use of midlevel practitioners to achieve labor cost savings in the primary care practice of an MCO. Health Services Research, 39(3), 607e626. Spetz, J., Parente, S. T., Town, R. J., & Bazarko, D. (2013). Scope-ofpractice laws for nurse practitioners limit cost savings that can be achieved in retail clinics. Health Affairs (Millwood), 32(11), 1977e1984. Stange, K. (2014). How does provider supply and regulation influence health care markets? Evidence from nurse practitioners and physician assistants. Journal of Health Economics, 33, 1e27. Swan, M., Ferguson, S., Chang, A., Larson, E., & Smaldone, A. (2015). Quality of primary care by advanced practice nurses: A systematic review. International Journal for Quality in Health Care, 27(5), 396e404. Thygeson, M., Van Vorst, K. A., Maciosek, M. V., & Solberg, L. (2008). Use and costs of care in retail clinics versus traditional care sites. Health Affairs (Millwood), 27(5), 1283e1292. US Census Bureau. (2013). Metropolitan and micropolitan statistical areas. Retrieved from http://www.census.gov/population/ metro/. US Department of Health and Human Services. (2013). Projecting the supply and demand for primary care practitioners through 2020. Retrieved from http://bhpr.hrsa.gov/ healthworkforce/supplydemand/usworkforce/ primarycare/. Virani, S. S., Maddox, T. M., Chan, P. S., Tang, F., Akeroyd, J. M., Risch, S. A., . Petersen, L. A. (2015). Provider type and quality
Nurs Outlook xxx (2016) 1e9
of outpatient cardiovascular disease care: Insights from the NCDR PINNACLE Registry. Journal of the American College of Cardiology, 66(16), 1803e1812. Watts, S. A., Gee, J., O’Day, M. E., Schaub, K., Lawrence, R., Aron, D., & Kirsh, S. (2009). Nurse practitioner-led multidisciplinary teams to improve chronic illness care: The
9
unique strengths of nurse practitioners applied to shared medical appointments/group visits. Journal of the American Academy of Nurse Practitioners, 21(3), 167e172. West, C. P., & Dupras, D. M. (2012). General medicine vs subspecialty career plans among internal medicine residents. JAMA, 308(21), 2241e2247.
Nurs Outlook xxx (2016) 1e9
www.nursingoutlook.org
Advanced practice clinicians as a usual source of care for adults in the United States Matthew A. Davis, MPH, PhDa,*, Cui Guo, MSb, Marita G. Titler, PhD, RN, FAANa, Christopher R. Friese, PhD, RN, AOCN, FAANa a
Department of Systems, Populations, and Leadership, University of Michigan School of Nursing, Ann Arbor, MI Department of Biostatistics Graduate Program, University of Michigan School of Public Health, Ann Arbor, MI
b
article info
abstract
Article history: Received 19 February 2016 Revised 22 June 2016 Accepted 4 July 2016
Background: Advanced practice clinicians (APCs) including nurse practitioners and
Keywords: Nurse practitioners Physician assistants Primary health care Delivery of health care Health costs
physician assistants are increasingly used to deliver care, yet little is known about these providers as a usual source of primary care. Purpose: This study examined the extent to which APCs serve as a usual source of care and the impact of such use on health care expenditures and quality. Methods: We performed a cross-sectional study by identifying 90,279 adults from the 2002 to 2013 Medical Expenditure Panel Survey who self-reported their usual source of care as either an APC or a primary care physician (PCP). Using complex survey design methods to make national estimates, we compared annual health care expenditures and quality measures among adults whose usual source of care is an APC to that of adults whose usual source of care is a PCP. Discussion: Nationally, 32 million adults visit an APC each year, yet only 1.4 million adults report their usual source of care to be an APC. In adjusted analyses, mean annual health care expenditures were $7,323 among APC patients vs. $7,959 among PCP patients, a difference of $635 (95% confidence interval [$1,408 to $138]). Across specific health services, APC patients trended toward having lower expenditures except for marginally higher expenditures on emergency room visits ($256 vs. $227 p < .001). APC patients were similar to that of PCP patients across health care quality measures. Conclusions: Few U.S. adults report their usual source of care to be an APC. Health care spending and quality measures are similar between APC patients and PCP patients. Expanding use of APCs as a usual source of care will likely not lead to increased health care spending. Cite this article: Davis, M. A., Guo, C., Titler, M. G., & Friese, C. R. (2016, -). Advanced practice clinicians as a usual source of care for adults in the United States. Nursing Outlook, -(-), 1-9. http://dx.doi.org/ 10.1016/j.outlook.2016.07.006.
Introduction As health care costs continue to rise (Keehan et al., 2015) and a shortage of primary care physicians
(PCPs) is evident (Bodenheimer & Pham, 2010; Ku, Jones, Shin, Bruen, & Hayes, 2011; West & Dupras, 2012), policy makers seek new strategies to meet the growing health care needs of the nation. The Patient Protection and Affordable Care Act contains several
* Corresponding author: Matthew A. Davis, University of Michigan, 400 North Ingalls, Room 4347, Ann Arbor, MI 48109. E-mail address: [email protected] (M.A. Davis). 0029-6554/$ - see front matter Ó 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.outlook.2016.07.006
2
Nurs Outlook xxx (2016) 1e9
provisions to improve access to care and expand health insurance coverage to millions of previously uninsured Americans. Upward of 7,200 additional physicians are required to provide primary care to the Americans who obtain coverage through the Affordable Care Act alone (Hofer, Abraham, & Moscovice, 2011; Huang & Finegold, 2013). Accounting for the aging population coupled with expansion of health insurance coverage, the U.S. Department of Health and Human Services projects a total shortfall of 20,400 PCPs by 2020 (US Department of Health and Human Services, 2013). Among others, the National Academy of Medicine advocates for increasing the use of advanced practice clinicians (APCs)dnurse practitioners and physician assistants to address the current and projected demand for primary care services (Health Affairs Health Policy Brief, 2012; Institute of Medicine, 2010; Kaiser Family Foundation, 2015; Naylor & Kurtzman, 2010; US Department of Health and Human Services, 2013). APCs currently represent a quarter of the U.S. primary care workforce and practice in diverse settings, which include physician-led health care teams, retail clinics, and as independent providers (Auerbach, 2012; Hooker, Brock, & Cook, 2016; Kralewski, Dowd, Curoe, Savage, & Tong, 2015; Roblin, Howard, Becker, Kathleen Adams, & Roberts, 2004; Stange, 2014; Thygeson, Van Vorst, Maciosek, & Solberg, 2008). Although they perform similar primary care services, insurers generally reimburse APCs at lower rates than physicians (Health Affairs Health Policy Brief, 2012; Naylor & Kurtzman, 2010; Stange, 2014), which may result in APC-delivered primary care costing less than that of PCP. However, lower direct costs associated with APC care may be offset by additional indirect costs because of greater reliance on diagnostic services (Hughes, Jiang, & Duszak, 2015), consultations to specialty physicians, and emergency department visits. In 2010, the National Academy of Medicine recommended that nurse practitioners (the largest constitute of the APC workforce) practice to the full extent of their training as independent clinicians (Institute of Medicine, 2010). However, to date, very little is known about the extent to which APCs practice as a usual source of care and the question remains whether APC-directed primary care differs in meaningful ways compared to PCP-directed care in regards to spending and quality. Therefore, we used nationally representative data from the Medical Expenditure Panel Survey (MEPS) to examine the use of APCs as a usual source of care in the United States. Our objectives were to (a) estimate the number of American adults who visit APCs and those who specifically self-report their usual source of care to be an APC, (b) compare the characteristics of U.S. adults whose usual source of care is an APC to that of adults whose usual source of care is a PCP, and (c) compare annual health care expenditures and quality of care among adults whose usual source of care is an APC to that of adults whose usual source of care is a PCP.
Methods Study Population and Data Source For this study, we pooled 2002 to 2013 data from the MEPS. Conducted annually by the Agency for Healthcare Research and Quality, the MEPS is a nationally representative sample of the noninstitutionalized U.S. population (Cohen, Cohen, & Banthin, 2009). Detailed information is gathered on health care utilization, expenditures, and health status. The MEPS utilizes an overlapping panel design consisting of household, medical provider, and insurance provider components. For each year, personal- and family-level data obtained from the household, medical provider, and insurance provider are collected and aggregated by the MEPS study team. For this study, we used data from the MEPS household component files including the full-year consolidated data and medical conditions files. This study was granted an exemption from institutional board review. We analyzed data from all adult (18 years and older) participants of the MEPS survey from 2002 to 2013. For the study time period, a total of 295,701 adults participated, with a range of 21,782 adults in 2007 to a high of 27,820 in 2012. From 2002 to 2013, the response rates among all participants ranged between 58.6% and 69.3%.
Measures Adults Who Visit APCs To determine how common APC care for adults is in the United States, we identified annual ambulatory visits to APCs (any visit to either a nurse practitioner or physician assistant in the calendar year) using data from the MEPS consolidated data files. From this, we estimated the total number of U.S. adults who use APC care each year.
Identification of Usual Source of Care Next, we used MEPS participants’ self-reported usual source of care as a surrogate measure of their primary care provider type (Friedberg, Hussey, & Schneider, 2010). MEPS participants are asked whether there is a particular doctor’s office, clinical, health center, or other place the individual usually goes to when sick or in need of health advice. From those who answered “yes,” we estimated the number of U.S. adults who have a usual source of care. For those who have a self-reported usual source of care, MEPS staff then inquires about whether their usual source of care is a “facility,” “person,” or “person within facility.” Those who report having a specific provider (i.e., either a person or person within a health care facility) are then asked to identify the professional credentials of the provider. We identified MEPS participants whose usual source of care was an APC
Nurs Outlook xxx (2016) 1e9
(defined as a “nurse practitioner,” or “physician assistant”) vs. PCP (defined as “MD, general/family practice” or “MD, internal medicine”). Our final analytic sample consisted of 90,279 adults who reported using either an APC or PCP as a usual source of care. Herein, “APC patients” refers to adults whose usual source of care is either a nurse practitioner or physician assistant, whereas “PCP patients” refers to adults whose usual source of care is a PCP.
Annual Health Care Expenditures In 6-month follow-up increments, MEPS participants are asked about all of their health care use such as ambulatory visits, inpatient and outpatient care, and prescription medications. MEPS staff then gathers information on the details regarding the health care use from providers and records how much was spent on the service. Aggregating across all health services, MEPS estimates the total amount spent in the year on health care servicesdwe used this as our primary end point. We extracted annual expenditure data on specific health services including office based, outpatient, prescription medications, emergency department visits, and inpatient care.
Health Care Quality Measures In addition to collecting information about health care expenditures, MEPS also captures a variety of selfreported health care quality measures. We adapted four primary care quality measures from the U.S. Health and Human Services core set of health care quality measures including the percentage of (a) adults between the ages 18 to 64 years who received an annual seasonal flu vaccination, (b) smokers who were encouraged by their provider to stop, (c) women between the ages 21 and 64 years who received a Pap smear examination within the past 3 years, and (d) diabetic patients who had an annual hemoglobin A1c (HbA1c) laboratory test (Kim, Park, Cha, & Jeong, 2015).
Characteristics of Study Participants We extracted sociodemographic data for participants in our study including age, sex, race/ethnicity (nonHispanic white, non-Hispanic black, Hispanic, vs. other), marital status (married, widowed/divorced/ separated, vs. never married), level of education (high school or less vs. above high school), U.S. region of residence (Northeast, Midwest, South, vs. West), health insurance coverage (private, public, vs. uninsured), and rurality. Metropolitan and micropolitan statistical areas (MSA; US Census Bureau, 2013) were used as a measure of ruralitydparticipants who resided in an MSA were considered nonrural vs. those residing in non-MSA areas were deemed to be in rural locales. As MEPS participants are surveyed multiple times over the calendar year, for measures that were acquired multiple times per year, we used the last measurement of the corresponding calendar year. As it is a strong predictor of health and mortality (DeSalvo, Bloser, Reynolds, He, & Muntner, 2006), we
3
used self-reported health status as an overall measure of health and collapsed it into “excellent, very good, or good” vs. “fair or poor.” Based on a combination of both physical and cognitive limitations measures, we also calculated the percentage of participants with “any functional limitation.” The MEPS also collects information on body mass index (BMI), smoking status, and administers the 12-item Short Form Survey (SF-12) to participants. We determined the percentage of participants who were obese (BMI 30 kg/m2) and the percentage who smoked, and we calculated the mean physical and mental summary component scores (PCS and MCS) that range from 0 to 100 (100 indicating the highest level of health). Finally, for each participant, based on all medical conditions coded by the International Classification of Diseases, version 9, Clinical Modification, we generated a comorbidity score using a modified version of the Charlson Comorbidity Index (Charlson, Pompei, Ales, & MacKenzie, 1987; D’Hoore, Bouckaert, & Tilquin, 1996).
Statistical Analyses Our primary analyses examined the relationship between usual source of care type (APC vs. PCP) and annual health care expenditures. Some study participants had no spending, and data from those with health care expenditures were highly skewed. Therefore, we used established modeling methods for zero inflated, skewed data that included a two-part multivariate model to generate predicted annual expenditures (Phillips et al., 2009). First, we used a logistic model to predict the probability of any spending. Second, among participants who had expenditures, we used a generalized linear model with log-link function and gamma distribution to predict annual expenditures. The final estimate of predicted annual spending was calculated by multiplying the predictions from the two models together. We did not use smearing methods because the residuals were heteroskedastic (Duan, 1983). For all analyses, we used complex survey design methods to make national estimates, which account for a participant’s probability of selection and sampling methodology. To increase the sample size of APC patients for analysis, we aggregated data across the 13 years and adjusted survey weights accordingly so that our final analytic data set represented 1 calendar year. To account for inflation, we used the Consumer Price Index from the U.S. Bureau of Labor Statistics for medical services to convert all expenditure data to 2013 dollars (Cavazos-Rehg et al., 2015). Our final models included covariates for sociodemographic differences, number of chronic conditions (based on the Charlson index), U.S. region (a surrogate measure of scope of APC practice [Health Affairs Health Policy Brief, 2012; Spetz, Parente, Town, & Bazarko, 2013]), and calendar year (to adjust for difference in practice over time). We used a t test to compare means and a c2 test to compare proportions between APC and PCP patient. Analyses were based on complete case analysis, and we
4
Nurs Outlook xxx (2016) 1e9
assumed any missing values to be missing completely at random. A two-sided p value of <.05 was considered statistically significant. All analyses were conducted using Stata MP, version 14.0 (College Station, TX).
Table 1 e Characteristics of U.S. Adults by Usual Source of Care Type Characteristic Usual Source p of Care Type Value* APC
Findings Nationally, we estimate approximately 32 million adults (14.2% of the 226 million U.S. adults) visit an APC each year. However, only 1.4 million adults (<1% among the 171 million adults who have a usual source of care) use APCs as a usual source of care, whereas 79.4 million use PCPs (Table 1).
Characteristics of APC vs. PCP Patients The characteristics of adults who report their usual source of care to be an APC differed in several ways from adults whose usual source of care is PCPs. Adult APC patients were younger, more likely to be female, and more likely to be non-Hispanic white than PCP patients ( p < .001 for all comparisons; Table 1). The percentage of adults residing in rural areas was more than doubled for APC patients compared to PCP patients ( p < .001). Finally, APC patients were more likely to be uninsured than PCP patients ( p < .001). APC patients were more likely to have poor or fair self-reported health status, to be obese, and to smoke than PCP patients (Figure 1). The mean PCS of the SF-12 was 47.4 (standard error [SE]: 0.52) among APC patients compared to 48.1 (SE: 0.08) among PCP patients ( p ¼ .20), whereas the mean MCS of the SF-12 was 49.8 (SE: 0.34) for APC patients compared to 51.2 (SE: 0.06) for PCP patients ( p < .001).
Annual Health Care Expenditures by Usual Source of Care Type In unadjusted analyses, the mean annual health care expenditures were $6,760 among patients whose usual source of care was an APC vs. $6,719 among patients whose usual source of care was a PCP, a difference of $41 (95% confidence interval [CI] [909 to 991]), p ¼ .93 (Table 2). Across specific expenditures, APC patients appeared to have equivalent or higher health care expenditures compared to PCP patients, albeit the only statistically significant difference was in annual expenditures on prescription medicationdAPC patients spent $337 (95% CI [58 to 615]), p ¼ .02, more per year on prescription medications than PCP patients. In adjusted models, mean annual health care expenditures were $7,323 among patients whose usual source of care was APC compared to $7,959 among patients whose usual source of care was a PCP, a difference of $635 (95% CI [1,408 to 138]), p ¼ .11. The only statistically significant differences between APC and PCP patients were expenditures on outpatient
Sample size, number 1,444 National estimates, millions (%)y Number of adults 1.4 Sociodemographic characteristics Age category (years) 18 to <45 0.6 (41.8) 45 to <65 0.6 (40.2) 65 0.3 (18.0) Sex Male 0.4 (31.6) Female 1.0 (68.4) Race Non-Hispanic white 1.2 (86.4) Non-Hispanic black 0.1 (5.6) Hispanic 0.1 (4.4) Other 0.0 (3.5) Marital status Married 0.8 (55.5) Widowed/divorced/ 0.3 (24.5) separated Never married 0.3 (20.0) Education High school or less 0.7 (48.9) education Above high school 0.7 (51.1) education U.S. census region Northeast 0.3 (19.7) Midwest 0.3 (20.6) South 0.5 (35.6) West 0.3 (24.1) MSA, rurality Non-MSA 0.4 (37.1) MSA 0.8 (62.9) Health care insurance Any private 1.0 (70.2) Public only 0.3 (20.5) Uninsured 0.1 (9.3)
PCP 88,835 79.4
<.001 29.8 (37.5) 30.5 (38.4) 19.1 (24.0) <.001 36.0 (45.4) 43.3 (54.6) <.001 59.8 (75.4) 7.8 (9.7) 6.9 (8.7) 4.9 (6.2) .14 47.0 (59.2) 17.7 (22.2) 14.7 (18.5) .11 35.6 (44.9) 43.7 (55.1) .17 19.3 (24.2) 17.3 (21.8) 29.0 (36.6) 13.8 (17.4) <.001 12.3 (15.6) 66.6 (84.4) <.001 61.1 (76.9) 13.7 (17.2) 4.6 (5.8)
Note. APC, advanced practice clinicians; MSA, metropolitan statistical area; PCP, primary care physician. * Difference compares APC vs. PCP patients. c2 test used to compare proportions. y Estimates based on weighted sample using complex design methods.
services and emergency department visits. Annually, APC patients spent $51 (95% CI [95 to 6]) less on outpatient services and $28 (95% CI [14 to 43]) more on emergency room visits than PCP patients.
Quality of Care Across the four measures of adult primary care quality, APC patients were comparable to PCP patients (Figure 2). The only marginal differences were a higher percentage of APC patients who reported receiving seasonal flu vaccination and diabetics who received an annual HbA1c compared to PCP patients ( p ¼ .10 and .07 respectively).
5
Nurs Outlook xxx (2016) 1e9
50 APC 45
PCP
< 0.001
40 35 30
0.28
Percent 25
< 0.01 0.04
20 15
0.46
10 5 0 1 Functional Poor or fair limitations health status
Obese
Smoker
2 Chronic conditions
Figure 1 e Health status of U.S. adults by a usual source of care type. APC, advanced practice clinician; PCP, primary care physician.
Discussion To our knowledge, this is the first national study to examine APCs as a usual source of adult care as compared to PCPs. We found that although 32 million adults visit APCs at least once during the year (14.2%),
100 90
only 1.4 million (<1%) report their usual source of care to be an APC. Despite the sizeable contribution of APCs to the primary care workforce and the 2010 National Academy of Medicine’s recommendation that nurse practitioners (the predominate APC provider) practice to their full independent potential (Institute of Medicine, 2010), we were surprised to find so few U.S.
0.07
APC PCP
0.29
80 0.37
70 60
Percent 50 0.10
40 30 20 10 0 Received seasonal flu immunization a
Smokers Women who had Pap Diabetics who had encouraged to stop smear within 3 years b annual Hb A1c c
Figure 2 e Patient-reported adult primary care quality measures by usual source of care type. aRestricted to adults aged between 18 and 64 years. bRestricted to women aged between 21 and 64 years. cRestricted to diabetics between 18 and 65 years. APC, advanced practice clinician; HbA1c, hemoglobin A1c; PCP, primary care physician.
6
Nurs Outlook xxx (2016) 1e9
Table 2 e Difference in Mean Annual Healthcare Expenditures by Usual Source of Care Type* Health Service
Unadjusted All health services Office-based care Outpatient services Prescription medications Emergency department visits Inpatient services Adjustedx All health services Office-based care Outpatient services Prescription medications Emergency department visits Inpatient services
Mean Expenditure by Usual Source of Care Type
Difference (95% CI)y
p Valuez
APC
PCP
$6,760 $1,598 $662 $1,929 $232 $1,938
$6,719 $1,564 $660 $1,592 $206 $1,870
$41 (909 to 991) $34 (192 to 261) $34 (192 to 260) $337 (58 to 615) $26 (11 to 62) $68 (765 to 901)
.93 .77 .97 .02 .17 .87
$7,323 $1,713 $645 $1,826 $256 $1,870
$7,959 $1,739 $695 $1,965 $227 $1,978
$635 (1,408 to 138) $26 (144 to 92) $51 (95 to 6) $139 (375 to 97) $28 (14 to 43) $108 (286 to 70)
.11 .67 .03 .25 <.001 .23
Note. 95% CI, 95% confidence interval; APC, advanced practice clinicians; PCP, primary care physician. * All estimates based on weighted sample using complex design methods. y Difference compares mean annual expenditures between APC vs. PCP patients (i.e., mean expenditurephysician mean expendituresAPC). z T test used to compare means. x Adjusted for age, sex, race/ethnicity, obesity status, smoking status, calendar year, U.S. region, and number of chronic conditions (Charlson index score).
adults reporting their usual source of care as an APC. Previous studies have found that access to primary care services is insensitive to variation in APC availability and more favorable nurse practitioner scope of practice (Cross & Kelly, 2015; Stange, 2014). This suggests that APCs primarily function as participants in collaborative primary care delivery models, rather than as independent providers who serve as an alternative portal of entry to the health care system. Another possibility is that Americans’ views regarding the use of APCs have not yet shifted to reflect their contribution as independent primary care providers, thus underestimating APCs’ potential contributions. As the scope of APC practice legislation expands across states, it may be worth examining public perceptions of APCs as usual sources of care. Our findings that patients whose usual source of care was an APC are more likely to be female, younger, uninsured, and of poorer health status are congruent with prior research (DesRoches et al., 2013). Similar to prior studies, we found that adults who use APCs were more likely to reside in rural areas suggesting APCs care for traditionally underserved populations (Everett, Schumacher, Wright, & Smith, 2009; Grumbach, Hart, Mertz, Coffman, & Palazzo, 2003; Larson, Palazzo, Berkowitz, Pirani, & Hart, 2003). Collectively, these findings support the National Academy of Medicine’s report that APCs care for vulnerable populations in rural areas where primary care services by PCPs are limited. Variation in access to primary care already exists across the country and 7 million Americans live in places where the demand for primary care services greatly exceeds supply (Huang & Finegold, 2013). Projected growth in the number of APCs may help offset
the demand for PCP services in these locales; however, variation exists in the supply of APCs as well (Kaiser Family Foundation, 2015; US Department of Health and Human Services, 2013). Increasing the availability of APCs through innovative educational programs is needed to meet these identified demands (Institute of Medicine, 2010). Patients whose usual source of care was an APC had comparable spending (if not lower) to those with a PCP as a usual source of caredthe exception being marginally lower annual outpatient expenditures and slightly higher expenditures on emergency room visits. A previous report found that health care spending varies by usual source of care with higher spending among adults who seek care from medical specialists (Phillips et al., 2009). Although not statistically significant ( p ¼ .11), in adjusted analyses, annual health care expenditures among adults whose usual source of care was an APC were $635 lower than those who see PCPs. This finding parallels a recent report of older adults that found those who sought most care from nurse practitioners had 29% less spending on evaluation and management compared to those who see medical physicians (Perloff, DesRoches, & Buerhaus, 2016). Such differences in spending may suggest large effects on national health care costs. Based on our estimate of the difference in annual spending among the 1.4 million adults who use APCs as a usual source of care, the net difference on national health care spending would be a reduction in approximately $890 million per year, however, it could range from a decrease of $2.0 billion to an increase in $193 million. APCs receive lower reimbursement than physicians for similar services (Health Affairs Health Policy Brief,
Nurs Outlook xxx (2016) 1e9
2012; Naylor & Kurtzman, 2010; Stange, 2014). Although not equitable, lower reimbursement among APCs may be the mechanism underpinning lower costs observed among APC-delivered care compared to that of PCPs (Martin-Misener et al., 2015). Prior studies do not account for potential indirect effects such as reliance on other services when compare APC to PCP care (Hughes et al., 2015). One advantage of our approach is that we were able to capture the expenditures on specific health services that suggest lower spending on prescription medications and outpatient services (many of which are diagnostic procedures such as imaging) but slightly higher spending on emergency department services. As these analyses were adjusted for patient characteristics, this suggests some differences in clinical practice styles between APCs and PCPs. Our findings that APC care is equivalent to that of physicians echo a large literature dating back to the 1980s (Naylor & Kurtzman, 2010; Newhouse et al., 2011; Swan, Ferguson, Chang, Larson, & Smaldone, 2015). Previous studies have investigated the management of specific diseases by APCs including diabetes and encouraging healthy behaviors such as smoking cessation (Dierick-van Daele, Metsemakers, Derckx, Spreeuwenberg, & Vrijhoef, 2009; Hughes et al., 2015; Kuo, Chen, Baillargeon, Raji, & Goodwin, 2015; Mundinger et al., 2000; Virani et al., 2015). Our study contributes to this body of work by finding that APC care appears to be comparable to that of PCPs in regard to current aggregate measures of health care quality based on U.S. estimates (Kim et al., 2015; Newhouse et al., 2011).
Limitations Our study has several limitations that must be acknowledged. First, the premise of our study was based on the assumption that the usual source of care oversees much of the care patients receive (Phillips et al., 2009). Therefore, our analyses did not directly account for the provider who ordered diagnostic tests or directed medical referrals. The potential mixing of provider types participating in care delivery would likely bias our results toward the null, hence making the estimates reported here conservative. Second, MEPS data are self-reported by participants, and therefore inaccuracies could have introduced measurement error. Nevertheless, in the assignment of patients to provider types and on quality measures, self-report is advantageous over other methods of assignment that rely on assumptions made by investigators. Third, the study population was limited to noninstitutionalized U.S. adult citizens; thus, findings among children or institutionalized adults may differ. Many APCs provide pediatric care, and our analyses did not examine associations among children. Fourth, because of the sample size of MEPS respondents who reported their usual source of care to be either a nurse practitioner or physician assistant, our definition of APC combined nurse practitioners and physician assistants. While consistent with recent approaches
7
(Hughes et al., 2015), our analyses are therefore not able to identify potential differences between nurse practitioners and physicians assistants. Finally, as our study is an observational cross-sectional design, we cannot rule out the potential impacts of residual confounding on our findings.
Future Recommendations Findings from this study and those of other investigators noted above provide a scientific basis for a call to action for state policy legislation that promotes APCs to practice to the full scope for which they are prepared. Second, findings support those of other investigators that APCs quality of care is equivalent to that delivered by PCPs. Given this equivalency in quality of care, it is concerning that the number of U.S. adults reporting APCs as the usual source of care is so low and calls for studies to explicate perceptions of U.S. adults regarding services and quality of care delivered by APCs in primary care. In addition, it is recommended that collaborations are fostered with selected patient advocacy groups to increase their understanding and value of services provided by these clinicians. The sociodemographic characteristics of patients cared for by APCs (rural, uninsured, and those with poorer health status, particularly mental health) suggest that these providers are filling a gap in health care, and at a lower or comparable cost to PCPs. In particular, there are innovative models of care delivery for adults with chronic conditions that include APCs in development. For instance, in Australia, for patients with chronic illness, nurse practitioners routinely play a key role in ongoing management (Watts et al., 2009). The findings from our study call for innovative and accelerated education of APCs who can meet the demands for primary care and whose educational trajectory is shorter than those for PCPs. To fill the projected gap in primary care of the future, federal and state policies must address the need for APCs to practice to their full potential and increase the funding of education for APCs.
references
Auerbach, D. I. (2012). Will the NP workforce grow in the future? New forecasts and implications for healthcare delivery. Medical Care, 50(7), 606e610. Bodenheimer, T., & Pham, H. H. (2010). Primary care: Current problems and proposed solutions. Health Affairs (Millwood), 29(5), 799e805. Cavazos-Rehg, P. A., Krauss, M., Fisher, S. L., Salyer, P., Grucza, R. A., & Bierut, L. J. (2015). Twitter chatter about marijuana. Journal of Adolescent Health, 56(2), 139e145. Charlson, M. E., Pompei, P., Ales, K. L., & MacKenzie, C. R. (1987). A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. Journal of Chronic Diseases, 40(5), 373e383. Cohen, J. W., Cohen, S. B., & Banthin, J. S. (2009). The Medical Expenditure Panel Survey: A national information resource to
8
Nurs Outlook xxx (2016) 1e9
support healthcare cost research and inform policy and practice. Medical Care, 47(7 Suppl 1), S44eS50. Cross, S., & Kelly, P. (2015). Access to care based on state nurse practitioner practice regulation: Secondary data analysis results in the Medicare population. Journal of the American Association of Nurse Practitioners, 27(1), 21e30. D’Hoore, W., Bouckaert, A., & Tilquin, C. (1996). Practical considerations on the use of the Charlson Comorbidity Index with administrative data bases. Journal of Clinical Epidemiology, 49(12), 1429e1433. DeSalvo, K. B., Bloser, N., Reynolds, K., He, J., & Muntner, P. (2006). Mortality prediction with a single general self-rated health question. A meta-analysis. Journal of General Internal Medicine, 21(3), 267e275. DesRoches, C. M., Gaudet, J., Perloff, J., Donelan, K., Iezzoni, L. I., & Buerhaus, P. (2013). Using Medicare data to assess nurse practitioner-provided care. Nursing Outlook, 61(6), 400e407. Dierick-van Daele, A. T., Metsemakers, J. F., Derckx, E. W., Spreeuwenberg, C., & Vrijhoef, H. J. (2009). Nurse practitioners substituting for general practitioners: Randomized controlled trial. Journal of Advanced Nursing, 65(2), 391e401. Duan, N. (1983). Smearing estimate: A nonparametric retransformation method. Journal of the American Statistical Association, 78(383), 605e610. Everett, C. M., Schumacher, J. R., Wright, A., & Smith, M. A. (2009). Physician assistants and nurse practitioners as a usual source of care. Journal of Rural Health, 25(4), 407e414. Friedberg, M. W., Hussey, P. S., & Schneider, E. C. (2010). Primary care: A critical review of the evidence on quality and costs of health care. Health Affairs (Millwood), 29(5), 766e772. Grumbach, K., Hart, L. G., Mertz, E., Coffman, J., & Palazzo, L. (2003). Who is caring for the underserved? A comparison of primary care physicians and nonphysician clinicians in California and Washington. Annals of Family Medicine, 1(2), 97e104. Health Policy Brief. (2012). Nurse practitioners and primary care. Health Affairs (Millwood). Retrieved from http://www. healthaffairs.org/healthpolicybriefs/brief.php?brief_id=79. Hofer, A. N., Abraham, J. M., & Moscovice, I. (2011). Expansion of coverage under the Patient Protection and Affordable Care Act and primary care utilization. Milbank Quarterly, 89(1), 69e89. Hooker, R. S., Brock, D. M., & Cook, M. L. (2016). Characteristics of nurse practitioners and physician assistants in the United States. Journal of the American Association of Nurse Practitioners, 28(1), 39e46. Huang, E. S., & Finegold, K. (2013). Seven million Americans live in areas where demand for primary care may exceed supply by more than 10 percent. Health Affairs (Millwood), 32(3), 614e621. Hughes, D. R., Jiang, M., & Duszak, R., Jr. (2015). A comparison of diagnostic imaging ordering patterns between advanced practice clinicians and primary care physicians following office-based evaluation and management visits. JAMA Internal Medicine, 175(1), 101e107. Institute of Medicine. (2010). The Future of Nursing: Leading Change, Advancing Health. Washington, DC: The National Academies Press. Kaiser Family Foundation. (2015). Issue brief: Tapping nurse practitioners to meet the rising demand for primary care. Retrieved from http://kff.org/medicaid/issue-brief/tapping-nursepractitioners-to-meet-rising-demand-for-primary-care/. Keehan, S. P., Cuckler, G. A., Sisko, A. M., Madison, A. J., Smith, S. D., Stone, D. A., & Lizonitz, J. M. (2015). National health expenditure projections, 2014-24: Spending growth faster than recent trends. Health Affairs (Millwood), 34(8), 1407e1417. Kim, H., Park, J., Cha, M., & Jeong, J. (2015). The effect of bad news and CEO apology of corporate on user responses in social media. PLoS One, 10(5), e0126358.
Kralewski, J., Dowd, B., Curoe, A., Savage, M., & Tong, J. (2015). The role of nurse practitioners in primary healthcare. American Journal of Managed Care, 21(6), e366ee371. Ku, L., Jones, K., Shin, P., Bruen, B., & Hayes, K. (2011). The states’ next challengeesecuring primary care for expanded Medicaid populations. New England Journal of Medicine, 364(6), 493e495. Kuo, Y. F., Chen, N. W., Baillargeon, J., Raji, M. A., & Goodwin, J. S. (2015). Potentially preventable hospitalizations in Medicare patients with diabetes: A comparison of primary care provided by nurse practitioners versus physicians. Medical Care, 53(9), 776e783. Larson, E. H., Palazzo, L., Berkowitz, B., Pirani, M. J., & Hart, L. G. (2003). The contribution of nurse practitioners and physician assistants to generalist care in Washington State. Health Services Research, 38(4), 1033e1050. Martin-Misener, R., Harbman, P., Donald, F., Reid, K., Kilpatrick, K. , Carter, N., . DiCenso, A. (2015). Cost-effectiveness of nurse practitioners in primary and specialised ambulatory care: Systematic review. BMJ Open, 5(6), e007167. Mundinger, M. O., Kane, R. L., Lenz, E. R., Totten, A. M., Tsai, W. Y., Cleary, P. D., . Shelanski, M. L. (2000). Primary care outcomes in patients treated by nurse practitioners or physicians: A randomized trial. JAMA, 283(1), 59e68. Naylor, M. D., & Kurtzman, E. T. (2010). The role of nurse practitioners in reinventing primary care. Health Affairs (Millwood), 29(5), 893e899. Newhouse, R. P., Stanik-Hutt, J., White, K. M., Johantgen, M., Bass, E. B., Zangaro, G., . Weiner, J. P. (2011). Advanced practice nurse outcomes 1990-2008: A systematic review. Nursing Economics, 29(5), 230e250. Perloff, J., DesRoches, C. M., & Buerhaus, P. (2016). Comparing the cost of care provided to Medicare beneficiaries assigned to primary care nurse practitioners and physicians. Health Services Research, 51(4), 1407e1423. Phillips, R. L., Dodoo, M. S., Green, L. A., Fryer, G. E., Bazemore, A. W., McCoy, K. I., & Petterson, S. M. (2009). Usual source of care: An important source of variation in health care spending. Health Affairs (Millwood), 28(2), 567e577. Roblin, D. W., Howard, D. H., Becker, E. R., Kathleen Adams, E., & Roberts, M. H. (2004). Use of midlevel practitioners to achieve labor cost savings in the primary care practice of an MCO. Health Services Research, 39(3), 607e626. Spetz, J., Parente, S. T., Town, R. J., & Bazarko, D. (2013). Scope-ofpractice laws for nurse practitioners limit cost savings that can be achieved in retail clinics. Health Affairs (Millwood), 32(11), 1977e1984. Stange, K. (2014). How does provider supply and regulation influence health care markets? Evidence from nurse practitioners and physician assistants. Journal of Health Economics, 33, 1e27. Swan, M., Ferguson, S., Chang, A., Larson, E., & Smaldone, A. (2015). Quality of primary care by advanced practice nurses: A systematic review. International Journal for Quality in Health Care, 27(5), 396e404. Thygeson, M., Van Vorst, K. A., Maciosek, M. V., & Solberg, L. (2008). Use and costs of care in retail clinics versus traditional care sites. Health Affairs (Millwood), 27(5), 1283e1292. US Census Bureau. (2013). Metropolitan and micropolitan statistical areas. Retrieved from http://www.census.gov/population/ metro/. US Department of Health and Human Services. (2013). Projecting the supply and demand for primary care practitioners through 2020. Retrieved from http://bhpr.hrsa.gov/ healthworkforce/supplydemand/usworkforce/ primarycare/. Virani, S. S., Maddox, T. M., Chan, P. S., Tang, F., Akeroyd, J. M., Risch, S. A., . Petersen, L. A. (2015). Provider type and quality
Nurs Outlook xxx (2016) 1e9
of outpatient cardiovascular disease care: Insights from the NCDR PINNACLE Registry. Journal of the American College of Cardiology, 66(16), 1803e1812. Watts, S. A., Gee, J., O’Day, M. E., Schaub, K., Lawrence, R., Aron, D., & Kirsh, S. (2009). Nurse practitioner-led multidisciplinary teams to improve chronic illness care: The
9
unique strengths of nurse practitioners applied to shared medical appointments/group visits. Journal of the American Academy of Nurse Practitioners, 21(3), 167e172. West, C. P., & Dupras, D. M. (2012). General medicine vs subspecialty career plans among internal medicine residents. JAMA, 308(21), 2241e2247.