PDF Archive

Easily share your PDF documents with your contacts, on the Web and Social Networks.

Share a file Manage my documents Convert Recover PDF Search Help Contact

Med Care Res Rev 2004 Lenz 332 51 .pdf

Original filename: Med Care Res Rev-2004-Lenz-332-51.pdf

This PDF 1.4 document has been generated by / Apogee Series3 Create v1.0.1, and has been sent on pdf-archive.com on 13/12/2015 at 17:16, from IP address 89.160.x.x. The current document download page has been viewed 345 times.
File size: 114 KB (20 pages).
Privacy: public file

Download original PDF file

Document preview

et al.61:3
/ Primary

Empirical Reseach

Primary Care Outcomes in
Patients Treated by Nurse Practitioners
or Physicians: Two-Year Follow-Up
Elizabeth R. Lenz

Ohio State University

Mary O’Neil Mundinger
Columbia University

Robert L. Kane

University of Minnesota

Sarah C. Hopkins

Columbia University

Susan X. Lin

Columbia University
This study reports results of the 2-year follow-up phase of a randomized study comparing
outcomes of patients assigned to a nurse practitioner or a physician primary care practice. In the sample of 406 adults, no differences were found between the groups in health
status, disease-specific physiologic measures, satisfaction or use of specialist, emergency
room or inpatient services. Physician patients averaged more primary care visits than
nurse practitioner patients. The results are consistent with the 6-month findings and
with a growing body of evidence that the quality of primary care delivered by nurse practitioners is equivalent to that by physicians.


comparative study; health services research; nurse practitioner; outcomes assessment; primary health care/standards

This article, submitted to Medical Care Research and Review on April 16, 2002, was revised and accepted for publication on June 5, 2003.
Medical Care Research and Review, Vol. 61 No. 3, (September 2004) 332-351
DOI: 10.1177/1077558704266821
© 2004 Sage Publications


Downloaded from mcr.sagepub.com at Vodafone on December 13, 2015

Lenz et al. / Primary Care Outcomes


During the past three decades, mounting evidence has supported the assertion that the quality of primary care delivered by nurse practitioners is equal
to that by physicians. The landmark 1974 Burlington randomized trial
assessed 1-year health outcomes of patients treated by nurse practitioners or
physicians and found the results to be comparable with respect to physical,
emotional, and social function; mortality; satisfaction; and quality of care
(Sackett et al. 1974; Spitzer et al. 1974). In 1979, Sox examined 40 evaluation
studies of nurse practitioner or physician assistant-delivered office-based care
and found their care to be indistinguishable from physician care. Simborg,
Starfield, and Horn (1978) conducted chart reviews of physicians and nurse
practitioners in six primary care practices and found nurse practitioners were
more likely than physicians to identify symptoms and signs in their patients,
to prescribe nondrug therapies, and to document follow-up of problems and
therapies. Patients attending a hypertension clinic were examined by Ramsay,
McKenzie, and Fish (1982) and were shown to have more success in managing
obesity and hypertension control if they were under the care of a nurse practitioner versus a physician. Attrition and appointment keeping, however, were
comparable among nurse practitioner and physician patients. Appointment
keeping among nurse practitioner and physician patients was also shown to
be comparable for nurse practitioners’ and physicians’ patients in a 1984
study, along with compliance, knowledge, recall of health counseling, and resolution of health problems. However, nurse practitioner patients were more
likely to be “completely” satisfied with their care (Powers, Jalowiec, and
Reichelt 1984). Using case vignettes posed to nurse practitioners and physicians, Avorn, Everitt, and Baker (1991) found that nurse practitioners were
more thorough in history taking and prescribed few medications. One retrospective review of nurse practitioner-physician comparative studies found
that nurse practitioners prescribe fewer drugs, order less expensive tests, and
use lower cost treatments, at comparable quality (Safriet 1992). Another
retrospective review found that nurse practitioners provide high-quality care
and that patients exhibit high satisfaction under their care (Carrino and
Garfield 1995).

The study was funded in part by the Robert Wood Johnson Foundation (#032806) and the United
Hospital Fund (#980202S). Study authors gratefully acknowledge the important contributions to
the study by Annette Totten, MPA. Address correspondence to Mary O’Neil Mundinger, Columbia University School of Nursing, 630 West 168th Street, New York, NY 10032; phone: (212) 3053582; fax: (212) 305-1116; e-mail: mm44@columbia.edu.

Downloaded from mcr.sagepub.com at Vodafone on December 13, 2015


MCR&R 61:3 (September 2004)

A 1995 meta-analysis indicated that randomized studies showed greater
compliance for nurse practitioner patients, while nonrandomized trials found
greater patient satisfaction and resolution of pathological conditions (Brown
and Grimes 1995). Three studies using National Ambulatory Medical Care
Survey data showed that nurse practitioners tended to perform fewer office
surgical procedures (Moody, Smith, and Glenn 1999) and provide more health
teaching, counseling, and emphasize therapeutic or preventive care (Moody,
Smith, and Glenn 1999; Aparasu and Hegge 2001; Hooker and McCaig 2001).
Nurse practitioners providing primary care to patients with diabetes have
been found to be more likely than physicians to perform certain tests and to
provide education about nutrition, weight, exercise, and medications (Lenz
et al. 2002). No differences were found, however, in patient outcomes.
Our recent study of 1,316 patients randomly assigned to either a nurse practitioner-run or a physician-run primary care practice revealed no clinically
significant differences in patient outcomes (health status, satisfaction, health
service utilization, and selected physiological indicators of chronic illness status) 6 months following the initial visit to the assigned provider (Mundinger
et al. 2000). While supplying compelling evidence for the quality of nurse
practitioner practice because of its methodological rigor and the equivalence
of the scope and authority of providers in the two types of practices, this study
addressed only relatively short-term outcomes. The decision to examine outcomes at 6 months post–initial visit was based on the transient nature of the
population served by the practices and the assumption that the measured outcomes would be sensitive to the short-term receipt of primary care. Nevertheless, a longer follow-up period was deemed desirable to ascertain whether differences in outcomes emerge after more sustained receipt of primary care
from the designated type of provider.
The purpose of the follow-up (Phase 2) study was to compare the outcomes
of patients randomly assigned to nurse practitioner and physician primary
care practices 2 years following their initial visit to the practice. The analysis
reported here includes the 406 patients who made an initial and at least one
follow-up visit to the practice to which they were randomized and who did
not receive primary care from any other medical center–affiliated practice
during the 2-year follow-up period. Consistent with Phase 1 of the study, the
following hypothesis was tested: in an ambulatory primary care environment
in which nurse practitioners have the same authority, responsibilities, productivity requirements, and patient population as physicians, the outcomes
(health status, satisfaction with care, utilization of health services, and
selected disease-specific clinical indicators) will not differ for the two
provider groups.

Downloaded from mcr.sagepub.com at Vodafone on December 13, 2015

Lenz et al. / Primary Care Outcomes


The conceptual framework for both original and follow-up studies was
based on a combination of Donabedian’s (1988) structure-process-outcome
model and conceptualizations and studies of professional socialization (e.g.,
Hall 1975, 1983; Hardy and Conway 1978). Donabedian’s model provided the
broad logical umbrella linking structural elements—including the background and education of the providers; the characteristics of patients served;
resources available; and the organizational context of policies, interactional
patterns, and expectations in which care is provided—to processes and outcomes of care. The import of the context for practice is underscored by
Moscovice’s (1978) finding that the practice setting exerted a stronger influence on patterns of primary care than did level of training when registered
nurses and family nurse practitioners were compared.
Professional socialization theory asserts that during the process of being
educated in a profession, the trainee is indoctrinated not only with the substantive knowledge that is needed to practice the profession but also its
worldview, norms, and values. The process of socialization involves both formal and informal interactions and is ongoing. During the basic phases of their
education, nurses and physicians are socialized to view patients, health, and
illness somewhat differently. Physicians, educated according to the medical
model, tend to adopt a more biological explanation of illness and to take a
system- or disease-specific orientation, whereas nurses tend to view the
patient from a biopsychosocial perspective, which emphasizes the whole
patient in a family and community context and stresses illness prevention.
However, nurses who go beyond their basic professional education and are
educated at the master’s level to be advanced practice nurses (for the purposes of this study, nurse practitioners) receive an additional overlay of powerful socializing experiences that are more akin to medical education in that
they focus on diagnosis and treatment of illness. Because nurse practitioners
and physicians receive many of the same socializing influences, it is reasonable to hypothesize that they would practice similarly and achieve similar
patient outcomes. The prediction of no differences was also underpinned by
the structural similarities between the two types of practices in the patient
populations served and the organizational context, which was identical.

Study participants were recruited into the study in 1995-1997 from the
emergency room and urgent care center of an urban academic medical center

Downloaded from mcr.sagepub.com at Vodafone on December 13, 2015


MCR&R 61:3 (September 2004)

hospital located in New York City in a predominantly Hispanic neighborhood. To be eligible, they must have presented with a nonemergent medical
condition and reported that they did not have a regular primary care provider
(PCP). If they agreed to participate, they were randomly assigned to primary
care follow-up at one of the medical center’s ambulatory care clinics (either a
nurse practitioner clinic or one of five physician clinics) and were given an
appointment with the designated provider. While their initial assignment was
to a specific provider within the nurse practitioner or physician practice,
patients were free to switch providers within the practice type to which they
were assigned.
The Phase 1 study sample was composed of the 1,316 patients who kept
their initial appointment with the PCP to whom they were assigned. Patient
outcomes were measured 6 months after the initial visit to the assigned provider. Results were that the two groups did not differ in health status at 6
months or in any of the health service utilization variables (primary care, specialist, or emergency room visits or hospitalizations) at 6 months or at 1 year.
The two groups did not differ in satisfaction following the initial visit, but
physician patients reported statistically higher average scores on one dimension of satisfaction (provider attributes) at 6 months. There were no nurse
practitioner-physician differences in either peak flow or glycosylated hemoglobin values at 6 months; however, diastolic blood pressure was significantly
lower in nurse practitioner patients. Both of the statistically significant differences were too small to be considered clinically significant. The conclusion
was that patient outcomes were comparable for nurse practitioner and
physician patients.

The goal at Phase 2 was to collect 2-year follow-up data from as many of the
Phase 1 participants as possible. The eligible sample for Phase 2 was composed of the 1,140 participants who were recruited into Phase 1 beginning January 1, 1996. (Research funding was secured for Phase 2 in 1997, at which
point the 176 patients who were recruited in 1995 had already passed the point
of their 2-year anniversary and thus were ineligible for follow-up.) Attempts
were made to contact all of the eligible participants by mail, then phone calls
and/or home visits were made to arrange and carry out data collection.
Of the eligible patients, 735 (65.7 percent) were located and interviewed for
Phase 2. A total of 405 participants were lost to follow-up. Of these, 88.4

Downloaded from mcr.sagepub.com at Vodafone on December 13, 2015

Lenz et al. / Primary Care Outcomes


percent could not be located, 1.2 percent were unable to complete the interview, and 10.4 percent refused. The response rate was consistent with that
reported for other studies conducted in similar populations (DeLia, Cantor,
and Sandman 2001; Marcus et al. 1998; Mickey et al. 1995; Pavlik et al. 1996).
Regarding the 735 participants who were interviewed, the mean age was 46.6
years, 80 percent were female, 92.6 percent were Hispanic, and 91 percent had
been enrolled in Medicaid at baseline.
According to medical center billing data, the 735 patients who completed
Phase 2 displayed four different patterns of primary health care receipt during
the 2 years following their initial visit to the practice to which they were
assigned: (1) 406 (55.2 percent) returned only to their originally assigned practice for primary care services after the initial visit, (2) 169 (23 percent) received
primary care from the assigned practice and at least one other practice, (3) 57
(7.8 percent) sought primary care only from other practices, and (4) 103 (14
percent) did not receive any additional primary health care from a medical
center practice. The latter participants gave reasons such as the following: not
being sick (51.8 percent), experiencing financial barriers to seeking care (18.5
percent), or being dissatisfied with their care (3.2 percent). The dissatisfied
participants were evenly distributed between nurse practitioner and
physician groups.
The present analysis is limited to the 406 patients who received primary
care only from the assigned practice and made at least one follow-up visit to
that practice during the 2 years following the initial visit. This subsample was
the only one that received the treatment as assigned and in which the effect of
the treatment could be isolated. Because the focus of the study was to compare
outcomes of patients receiving primary care from a nurse practitioner or a
physician, it seemed prudent to eliminate patients who had, in essence, contaminated the treatment by crossing over from one type of PCP to another. To
have included all participants in an intention to treat analysis would have provided a more conservative estimate of the effect of the intervention and thus
would have increased the likelihood of finding no difference. Because we
were testing a hypothesis of no difference, we did not want to employ any
device that would increase the chance of such a finding.
Demographics and baseline summary SF-36 scores were compared for the
406 patients to the remaining patients who did not receive the sustained intervention with (1) those who received primary care from other providers, either
from the assigned practice in addition to another practice or only another
practice (n = 226); (2) those who did not receive any additional primary care
(n = 103); and (3) those lost to follow-up (n = 405). Chi-square analysis and ttests revealed that the 406 patients who consistently stayed with their originally assigned practice (the present sample) were less likely to be diabetic (11.6

Downloaded from mcr.sagepub.com at Vodafone on December 13, 2015


MCR&R 61:3 (September 2004)

percent vs. 18.1 percent, p = .022) and had lower mental health scores (39.37 vs.
41.72, p = .042) when compared with those who sought primary care from
other sources. When compared with those who sought no primary care, the
406 patients who stayed with their original practice were significantly older
(mean age 46.51 vs. 41.12, p = .001), more likely to be enrolled in Medicaid (91.1
percent vs. 81.6 percent, p = .005), and to be hypertensive (37.4 percent vs. 24.3
percent, p = .012). The patients lost to follow-up were more likely to be male
(29.1 percent vs. 18.2 percent, p = .000) and not enrolled in Medicaid at baseline
(17.8 percent vs. 8.9 percent, p = .000). Chi-square analyses and t-tests were run
to compare the nurse practitioner patients lost to follow-up (n = 237) with the
physician patients lost to follow-up (n = 168). No significant differences were
found with respect to baseline demographics or self-reported health status.
Chi-square analyses and t-tests were also run to examine potential differences
between the nurse practitioner and physician patients who were interviewed
at Phase 2 but did not stay with their originally assigned provider. No significant differences were found for their baseline health status or most demographics, although nurse practitioner patients were slightly more likely to be
Hispanic (94.3 percent vs. 88.9 percent, p = .045). The most consistent pattern
across the above comparisons was the expected result that patients with
health care insurance coverage were more likely than those without to continue seeking primary health care during the 2-year period; in analyses of
health care utilization in which it was a highly relevant variable, insurance
status was controlled statistically.
For the analysis of patient satisfaction, the sample was further limited to the
217 patients who had received primary care from the assigned provider during year 2 (12 to 24 months following the initial visit). The reason for this limitation was the validity of the data generated from the standardized instrument. Instructions were that the participant was to refer only to the previous
year when answering the questions. The decision to limit the sample to only
those patients who received the pure intervention during the previous year
unfortunately resulted in a low statistical power to detect differences between
the groups and increased the likelihood that the null hypothesis would be
Participants were permitted to choose the site of data collection: either in
their homes, in a university office setting, or by phone. Data collection by
bilingual interviewers consisted of the verbal administration of the Medical
Outcomes Study Short–Form 36 (MOS SF-36) (Ware et al. 1993) to measure
health status and the Primary Care Assessment Survey (PCAS; Safran et al.

Downloaded from mcr.sagepub.com at Vodafone on December 13, 2015

Lenz et al. / Primary Care Outcomes


1998) to measure patient satisfaction. Subscales of the PCAS include Financial
Access, Organizational Access, Visit-Based Continuity, Knowledge of Patient,
Preventive Counseling, Integration, Interpersonal Treatment, Communication, and Trust. Scores are calculated for each subscale and are extrapolated to
a range from 0 to 100. The developers of the PCAS have found each of the
subscales to exhibit consistently strong reliability and validity.
Bilingual research nurses took blood pressures, peak flow readings, and
drew blood to measure glycosylated hemoglobin for the hypertensive, asthmatic, and diabetic patients, respectively. The medical center provided billing
data about participants’ use of primary, specialist, emergency room, and hospital care for the 2-year period following the initial visit to the assigned PCP
and 6 months prior to baseline. Data were analyzed using chi-square, t-tests,
and repeated measures analysis of variance (SPSS-PC).
Characteristics of Study Sample
The 406 patients in the study sample for Phase 2 were largely middle-aged
(mean age of 46.5 years), female (81.8 percent), Hispanic (92.3 percent), Spanish speaking (87.4 percent), and enrolled in Medicaid at baseline (91.1 percent). Fifty-nine percent of the participants had one or more of the targeted
chronic conditions: hypertension, diabetes, and/or asthma. Nurse practitioner and physician patient groups did not differ in age, gender, ethnicity, or target conditions (Table 1). The only difference between the two groups was in
the percentage who were enrolled in Medicaid at baseline, with patients in the
physician group more likely to be enrolled.
Health Status
Consistent with the findings of Phase 1, there were no statistically significant differences between nurse practitioner and physician patients in their
self-reported health status at 2 years, when measured with either the SF-36
subscales or the physical and mental health summary scores (Table 2). The
study sample had mean baseline scores on the SF-36 summary scores that fell
within the lowest quartile of national norms for the instrument (Ware et al.
Repeated measures analysis of variance was used to examine SF-36
subscale and summary score changes from baseline to 6 months and 2 years in
the study sample of 406 participants. Within-subjects (time-related) effects
were found for all subscales and summary scores (F values ranged from 12.11

Downloaded from mcr.sagepub.com at Vodafone on December 13, 2015


MCR&R 61:3 (September 2004)

TABLE 1 Demographic Characteristics of Nurse Practitioner and
Physician Groups (N = 406)
Nurse Practitioner
(n = 222)

Number Percentage


Target conditions
Medicaid status at baseline
Not enrolled

(n = 184)

Number Percentage







χ = 0.27






χ = 2.004






χ = 6.255






χ = .641
χ = .708
χ = .001






χ = 8.505







to 46.48, and in all cases p < .001). Within-subjects contrasts revealed that significant improvements occurred in all subscales and summary scores from
baseline to 6 months and from baseline to 2 years. These improvements were
expected because baseline data were collected just after an ER or urgent care
visit. Significant changes from 6 months to 2 years were found only for two
subscales: Social Functioning and Mental Health, both of which decreased.
There were no significant group-by-time interaction effects, suggesting that
patterns of change were similar in the nurse practitioner and physician
When the physiological indicators of nurse practitioner and physician
patients who were hypertensive, diabetic, or asthmatic were compared at 2
years following the initial visit, no significant differences were discerned
between the nurse practitioner and physician groups in any of the measures
(Table 3). It is noted that the glycosylated hemoglobin levels for the nurse
practitioner and physician groups indicate that in neither group was diabetes

Downloaded from mcr.sagepub.com at Vodafone on December 13, 2015

Related documents

med care res rev 2004 lenz 332 51
clinical study prescriber drug dispensing
medical assistant overview1097
medically unexplained symptoms oct 2016
cpg management of thalassaemia

Related keywords