Your Subtitle text

PA-MD Study

Physician assistants in occupational medicine:

how do they compare to occupational


R. S. Hooker

Background Physician assistants (PAs) have been present in occupational and environmental

medicine (OEM) in the USA since 1971, yet remarkably little is known about their


Methods An administrative study of PA activities was undertaken and compared with the

activities of physicians in the same occupational medicine setting. Patients were not

triaged to either provider and all resources of care were recorded for the visit. An

episode of care approach was used for the analysis.

Results The characteristics of patients seen by each provider were similar in age, gender ratio

and severity of injury. Physicians saw a mean of 2.9 patients/h and PAs 2.5, but

PAs worked more hours and saw more patients per year than physicians. The

average charge per patient visit and total charge for an episode of care were similar.

Differences between PAs and physicians were seen in the areas of ‘limited duty’

duration given to patients and on average PAs prescribed 15 days and physicians

17 days. PAs referred a patient 19.7% of the time, while physicians referred 17.4%.

Most of the referrals were to physical therapy. The salary of a physician, based on an

hourly rate, was approximately twice as much as a PA.

Conclusion The use of PAs in OEM may represent a cost-effective advantage from an

administrative standpoint. Clearly, more research is necessary in determining the

role and utilization of PAs in OEM and how they may improve the delivery of

physician services.

Key words Administration; cost analysis; industrial medicine; management; occupational

medicine; physician assistant; USA; utilization.

Received 3 December 2002

Revised 28 May 2003

Accepted 23 July 2003


The concept of occupational and environmental

medicine (OEM) has its origins with the first doctor

who administered medical attention to patients with a

work-related injury or illness. Since that time, the practice

of OEM has evolved, keeping pace with the needs of

society and the modern day workforce. In the USA,

the evolution of medicine has incorporated physician

assistants (PAs) to the medical team since 1967 [1]. PAs

are a type of medical provider with a unique set of skills

and licenced by all 50 states to be able to assess, manage

and treat patients, including prescribing of medication.

They are delegated these tasks by physicians and the care

they provide parallels that of physicians for the most part

[2]. Their training is comprehensive, lasting on average

26 months and largely parallels a medical school

curriculum, only in a briefer amount of time. There are

over 45 000 PAs in clinical practice in the US and

Canada. PA students enter an education program with

24 months or more, on average, of health care experience

in some medical field such as emergency medicine,

pharmacy, respiratory therapy, nursing, or medical

VA North Texas Healthcare System, Medical Service, 4500 S. Lancaster Road

(111), Dallas,TX 75216-7191, USA. Tel: +1 214/857 1544;


Occupational Medicine 2004;54:153–158

DOI: 10.1093/occmed/kqg126

Occupational Medicine, Vol. 54 No. 3

Published by Oxford University Press 153

Downloaded from by guest on February 13, 2012

corpsman [2]. Expansion of the PA concept to England

and The Netherlands began in 2002 [3].

In 1999, the American Academy of Physician Assistants

identified ~38 000 clinically active PAs and 500 were

employed in OEM [4]. While there are PAs in OEM in

all 50 states, what they do and how well they do it is

unknown. The survey instrument used by the American

Academy of Physician Assistants is a broad-based questionnaire

to gather information about PAs overall—

information about specialties is not detailed.

The introduction of PAs to OEM began with oil and

gas production expansion in the 1970s. This industry was

in need of health professionals for their growing employee

corps. Physicians were in short supply and corporations

turned to the growing source of PAs to help fill this void.

PAs were employed as early as 1971 on the Alaskan

pipeline, a huge enterprise that stretched from the Arctic

Ocean to an ice-free port in the southern part of the state

[5,6]. Also in the 1970s, PA educators began looking to

industrial medical settings as clinical rotation sites for

students [7]. The delivery of routine physical examinations

for industries and insurance companies by PAs

was seen as an economical strategy. The cost of a PA in

this service was considerably less than a physician [8].

Industrial studies conducted primarily by PAs instead of

physicians have been a trend in certain sectors since the

1980s [9,10].

In spite of the growing literature, very little information

exists concerning the role of PAs in the practice of

contemporary OEM. The few studies that have been

published are both limited in scope and dated in terms of

the evolution of OEM PAs. One report estimated that

some 500 PAs were providing OEM services directly to

the employees of 15 Fortune 500 companies [9]. Elliott

[9] reported that 58% of the 124 OEM PAs surveyed in

1983 indicated that they were employed directly by the

company and provided medical service at the plant site.

Another 15% indicated corporate medical staff positions

where they reported directly to the corporate medical

director. Overall, the survey respondents estimated that

55% of their time was spent providing direct patient care.

Additionally, 27% of respondents indicated they worked

in private industrial medicine clinics and were employed

by a physician or group of physicians rather than a

company. Ramos observed that many were providing

essentially primary care services along with other duties

[11]. A list of job descriptions of PAs in OEM is cited

elsewhere [10].

Since the early descriptions of PAs in OEM, many

changes have occurred which have enhanced and

modified the delivery of health care. Organizations are

undertaking cost-effective and administrative strategies

that are designed to improve the efficiency of their operation.

Although PAs would probably not be employed in

various OEM settings if they were not cost-effective, no

systematic analysis or comparative role study has been

undertaken to date.

Accordingly, a study of OEM PAs in one occupational

medicine organization was developed to better understand

some of the similarities and differences in provider

roles. We asked one fundamental question: how do PAs

and physicians in this organization compare in their

management of OEM patients?


This is primarily an administrative data study and not

medical outcomes research. Types of patients seen and

visit dispositions between two types of providers were

compared. The information sources were the billing

and other internal administrative datasets from a large

employer of PAs and physicians in OEM. Statistical

analyses were conducted and the 95% confidence intervals

around the means and proportions were computed

using SAS (Statistical Analysis System), a statistical

software package used in the biomedical sciences.

Differences between the numbers of observations for

each group of clinicians were investigated using analysis

of variance. Chi-square analyses determined differences

in proportion of treatment mentioned by type of provider

and other variables.


The setting was the Dallas–Fort Worth (DFW) Texas

region of Concentra Managed Care. Concentra is a

for-profit health care organization that specializes in

occupational and industrial health care and injury

treatment services. It is the largest OEM corporation in

North America and employs 8800 people in 32 states and

Canada. One aspect of this service is to manage the

disability costs associated with workers’ compensation

injuries. In 1999 the organization served ~130 000

different employers and processed >5 million patient


The Concentra presence in the DFW area includes 36

health care providers specializing in family practice,

OEM, general surgery, emergency medicine and physiatry.

Providers are defined as allopathic and osteopathic

physicians (MD/DOs) and PAs. In 1999, there were

~84 000 visits at 16 proprietary medical offices. A broad

range of services included injury management, employment

related physical examinations, pre-placement

substance abuse testing, regulated examinations (i.e.

respirator mask fittings), job-specific return-to-work

evaluations and information management. Specialized

‘on-site’ occupational health care and consulting services

were also performed.

Patients are referred from their employers to Concentra,

where they are assigned an appointment time and

location. Urgent cases are seen on the same day. Referrals


Downloaded from by guest on February 13, 2012

and return visit dates are assigned at the time of the visit.

There is no triage and patients are assigned to the next

available provider. Protocols are system wide and not

specific for a type of provider. A supervising physician is

administratively assigned to a PA for the day, as required

by Texas medical licensing statute.

Data sources

The data for this study include three sources: employee

administrative files maintained by the Department of

Human Resources; patient encounter files; and billing

records. For each encounter or patient visit, the provider

of record enters the information about the patient on

forms and in networked computers. Data input

employees further abstract the patient encounter forms

and add administrative data for billing purposes. These

data are warehoused on a mainframe computer at the

national headquarters in Dallas,Texas.

The patient encounter database contains a case or

episode of injury information for an individual. Each visit

can include a wide range of information for the

encounter. Important fields for this analysis include

patient demographics, date of injury, unique case record

number used for managing the encounter, number of

visits, treatment, referrals, medications prescribed, other

resources and the provider of record. This information is

then linked with the provider such as specialty, type, date

of hire, hours worked and other characteristics.

The activity of physicians and PAs in OEM was

examined using episode of care methodology. An episode

of care began on the date the employee sought care from

Concentra and ended when the case was closed to further

care. A diagnostic code (or codes) was entered at the first

encounter. All activity was linked to the patient and the

first encounter date. The diagnostic code could change,

but the date of the encounter and patient did not. Because

the intent of the study was to compare total care by two

types of providers if care was divided between two OEM

providers, or more than one diagnosis recorded, that

episode was eliminated from the analysis.

There were 168 000 patient visits to eight DFW

Concentra Managed Care medical offices in 1999. For

this study, the activities of interest were injuries, which

constitute 94% of all doctor and PA provided services at

Concentra Managed Care.


When this study was conducted in 1999, the DFW region

of Concentra employed 36 providers specializing in OEM

(Table 1). Two-thirds (24) were physicians and the other

third (12) were PAs. No nurse practitioners were

employed in this region at the time of this study. The

average age of the physician was 50 years and that of a PA

was 45 years. The mean annual salary was $143 056 for a

physician and $74 208 for a PA. Benefits were not

calculated in the salary structure. The mean total hours

worked for the physician was 1662.2 annually and

1871.3 for the PA. Annual hours included part-time and

overtime employment.

There were 29 676 first encounters and 51 088

follow-up visits available for analysis where one provider

managed an episode of care for all visits. How each group

of providers managed patients is displayed in Table 2. The

average number of days of ‘limited activity’ assigned to

the patient (as a form of injury management) by all

providers was 16.8, with PAs averaging 15.6 and

physicians 17.4 (P = 0.015). Referring a patient to an

outside provider (e.g. ENT surgeon) was rated as a

probability with an average likelihood of 18.2% of all

visits. Physicians were less likely (17.4%) than a PA

to refer (19.7%; P = 0.0001). An overall average of

2.8 patients/h was maintained in 1999: physicians

averaged 2.9 and PAs 2.5 (P = 0.008).The average charge

per visit was $296.72. The average total charge per

episode was $594.33. While visit charges for PAs were less

than physician charges, on average the difference was

negligible and did not reach statistical significance.

Providers used a three-point scale, the Injury Severity

Scale, at the time of encounter to assess the severity of

the patient’s complaint. Selected criteria categorized the

patient’s condition as mild, moderate, or severe. The

mean severity scale for both providers was 1.93.

Most of the patients were male (average 73.1%). While

the difference (72.5 and 74.3% for physicians and PAs,

respectively, P = 0.007) reached statistical significance,

the clinical difference was considered irrelevant.

The probability of patients being likely to keep their

appointments was assessed. The overall average was 79%,

with 81% for PAs and 76% for physicians (P = 0.002).

Finally, a comparison regarding types of referrals made

by each provider was conducted (Table 3). No statistical

Table 1. Occupational and environmental providers by labor characteristics (1999)

No. Age


Gender Mean salary


Average no. of

visits per hour

No. of hours



hours worked

per provider

Physicians 24 50 18 M:6 F 143 056 2.9 39 892 1662.2

Physician assistants 12 45 7 M:5 F 74 208 2.5 22 455.5 1871.3


Downloaded from by guest on February 13, 2012

difference emerged between physicians and PAs where

the referrals were directed.


In this particular OEM setting, some administrative

differences emerged between the two providers. PAs work

proportionally more hours on average than physicians

and their salary is approximately half of a physician’s

salary. This salary difference is consistent with nationally

observed PA and physician salaries in primary care [12].

Characteristics of patients are important variables

when comparing provider activities. In this study, no

significant differences emerged. The average age of

patients was similar between the two providers and the

percentage of males differed by only 2.2%. The injury

severity scale scores were identical. However, a threepoint

interval scale was not considered sensitive to

analysis and has not been validated. An interesting

observation was that patients seen by PAs were more

likely to keep their return appointments than patients

seen by physicians. One interpretation of this finding

might be a measure of compliance and could represent

satisfaction with care or some other attribute of either

the physician or PA. Patients who do not keep their

appointments also represent unrecompensed time.

The amount of revenue a provider generates is reflected

in the billing charges. In this setting, all OEM visits were

billable, therefore a higher rate of return visits to see the

PA may be viewed as beneficial to the organization,

especially since the charges were fixed regardless of

provider. The average charge for an unscheduled (injury)

visit was $296.72 (1999 dollars). An episode of care (two

visits for 93% of all visits) was $594.33 on average. The

differences in average charges generated by each group of

provider for a visit or an episode did not reach statistical


The management of patients revealed a few differences

between providers. When light duty or limited work

activity was assigned to an industrial medicine patient,

PAs prescribed less time (15.6 days) than physicians (17.4

days). The implications of these findings are beyond the

scope of this paper, since treatment based on diagnosis

was not part of this study. However, since patients were

undifferentiated (not triaged) to one type of provider

over another, it implies there may be differences in care

between providers.

Another area of interest in this study is the referral rate.

The average referral rate of 18.2% is 2.5 times as high

as the national rate in adult primary care [13,14]. This

finding may reflect that not all of OEM is primary care;

the need to use allied health (e.g. physical therapy) and

select specialists (e.g. orthopedics) for industrial patient

management may be higher than primary care.

While barriers to referral can limit the cost-effectiveness

of PAs [15], the referral rate of almost 20% by PAs in

this study suggests referral barriers were not an issue.

When the referrals were systematically analyzed, most of

the referrals were to physical therapy (PT) for both

providers. Some evidence suggests that a referral to PT

can accelerate recovery from an acute care mode and

return him or her to work more quickly [16]. Patients

seen by PAs for an episode of care were assigned shorter

periods of limited activity on average, but at a higher PT

referral rate than the patients of physicians, suggesting

there may be differences in how the patients are managed.

This was an administrative study designed to be useful

to managers of health providers and not an econometric

study. As such, there are some limitations to this type of

Table 2. Comparison of PAs and physicians by outcomes of episodes of care







P SD 95% CI

1. Average number of days of limited activity assigned 15.6 17.4 16.8 0.015 48.2 16.1–17.5

2. Likely to refer a patient to an outside provider 19.7% 17.4% 18.2% 0.0001 38.6% 17.6–18.7%

3. Average number of patient visits per hour 2.5 2.9 2.8 0.008 1.9 2.8–2.8

4. Average charge per visit $284.77 $302.53 $296.72 n.s. $159.76 $294.50–298.93

5. Average total charges for episode of care $565.98 $608.13 $594.33 n.s. $751.46 $583.91–604.75

6. Average severity score of problems treated

(mild = 1; moderate = 2; severe = 3)

1.92 1.93 1.93 n.s. 0.33 1.93–1.94

7. Percentage of male patients 74.3% 72.5% 73.1% 0.007 44.3% 72.5–73.2%

8. Average age of patients (years) 35.3 35.5 35.5 n.s. 11.2 35.3–35.7

9. Probability patients likely to keep their appointment 81% 76% 79% 0.0024 1.04 78.0–80.9

Table 3. Percentage distribution of referrals by type of provider


Referral type Physician PA

Physical therapy 84.0 88.0

Orthopedics 4.6 3.0

General surgery 2.1 2.3

Other 9.3 6.7

Total 100 100


Downloaded from by guest on February 13, 2012

work from a labor cost-effectiveness standpoint. As with

many studies of this type, it was a retrospective study

using secondary data for analysis. Data collected and

intended for one purpose may have unknown biases that

skew the data in one direction or another. If PA care

is comparable to a physician’s care in OEM, then a

prospective study that matches patients for age, gender,

type of injury, diagnosis, severity, time of visit and

resources used needs to be undertaken to substantiate

that belief.

Another limitation with this study is the unknown

interaction between the PA and physician.Technically, the

role of the PA in relation to the physician is one of

delegated responsibility. How the PA provides care can

theoretically be enhanced or inhibited by the physician’s

supervisory influence. This team role is identified as one

issue in the provision of patient care and needs to be

assessed from an overall productivity standpoint [17].

It is difficult to generalize these findings to OEM PAs

in other settings, since so little is known about OEM

PAs. Concentra is the largest OEM company in North

America, unique in its size, and there is no other company

that comes close. It may have an economy of scale that

makes comparisons difficult. However, the method it

employs to process patient needs and its use of PAs is

consistent with other managed care firms and probably

reflects contemporary organizational practices [18].While

each of the states differ somewhat in permissive legislation

concerning how PAs may perform their delegated

roles, the incorporation of employees in managed care

settings tends to dampen these differences and allows PAs

and other employees to function fairly efficiently. For

example, PAs in Texas, at the time of this study, could not

prescribe narcotic analgesics. However, policies within

the organization permitted PAs to phone in a controlled

substance in the supervising physician’s name thus

avoiding interruptions in workflow.


This small study offers some insight into the activity of

PAs in OEM, at least in the USA. In this particular

setting, where OEM is the exclusive focus of medical

care, it appears that PAs are providing a service that is

comparable to physician services from an administrative

viewpoint. Their productivity to compensation ratio

suggests they may be economical members of the health

care team from a labor standpoint. However, some of

their cost-effectiveness may be negated by a higher

referral rate than the physicians.This observation requires

further analysis to see if the benefit of a PA in the provider

mix is retained.

While it is probably safe to say that PAs would not be

employed in OEMif they were not cost-effective, far more

needs to be understood about PA delivered services from

an economic perspective. Clearly, little is known about

what OEM PAs do,where they work and how they benefit

their employers. Understanding how they are utilized

precludes any organizational evaluations that would

improve patient and employer services. It is hoped this

study helps to establish a foundation for additional

research on how OEM services can be effectively and

efficiently delivered.


The author gratefully acknowledges Sandi Mackey, PA,MPAS

and Robert Hassett, DO, who helped collect the data and

Charles McDonnell, PhD, who assisted with the data analysis.

The author also acknowledges James Davidson, PA, Mary Ann

Ramos, PA, Donna Lux Hooker, as well as the anonymous

reviewers of the paper for their suggestions.


1. Hooker RS, Cawley JF. Physician Assistants in American

Medicine, 2nd edn. New York: Churchill Livingstone, 2003.

2. Hooker RS, McCaig LF. Use of physician assistants and

nurse practitioners in primary care, 1995–1999. Health Aff

(Millwood) 2001;20:231–238.

3. Cawley JF, Hooker RS. Physician assistants: does the US

experience have anything to offer other countries? J Health

Services Res Policy 2003;8:65–67.

4. American Academy of Physician Assistants. Marvelle K.

1999 AAPA Census Report. Alexandria, VA: American

Academy of Physician Assistants, 1999.

5. Randall GB. Don’t Call Me Doctor. Kearney, NE: Morris,


6. Randall GB. A Medex in bush Alaska. Physician Assist J


7. Lynch J. Allied health personnel in occupational medicine:

report of the long range planning committee. J Occup Med


8. Weisenberger BL. Occupational medicine’s new resource.

J Occup Med 1974;16:676–677.

9. Elliott CH. The physician assistant—newest member of the

corporate health care team. A guide to hiring a PA and

cutting health care costs. Pers Admin 1984;29:87–92.

10. Ramos M. Occupational and environmental medicine. In:

Ballweg RM, Stolberg S, Sullivan E, eds. PhysicianAssistant:

A Guide to Clinical Practice, 3rd edn. Philadelphia, PA:

W. B. Saunders, 2003; 844–873.

11. Ramos M. Occupational medicine. An overview for

physician assistants. Physician Assist 1989;13:79–86.

12. Hooker RS. The economics basis of physician assistant

practice. Physician Assist 2000;24:51–71.

13. Bachman KH, Freeborn DK. HMO physicians’ use of

referrals. Soc Sci Med 1999;48:547–557.

14. Cherry DK, Burt CW,Woodwell DA. National Ambulatory

Medical Care Survey: 1999 Summary, 322. Advance Data

from Vital and Health Statistics. Hyattsville, MD: National

Center for Health Statistics, 2001; 1–36.

15. Enns SM, Muma RD, Lary MJ. Examining referral


Downloaded from by guest on February 13, 2012

practices of primary care physician assistants. J Am Acad

Physician Assist 2000;13:81–118.

16. Saladin LK, Morrisette DC, Brotherton SS. Making the

physical therapy referral. J Am Acad Physician Assist


17. Geller JM, Weier A, Muus KJ, Hart LG. Understanding

the interactions between physician assistants and their

supervising physicians in the provision of patient care.

J Allied Health 1998;27:188–195.

18. Freeborn DK, Pope CR. Promise and Performance in

Managed Care:The Prepaid Group Practice Model. Baltimore,

MD: Johns Hopkins University Press, 1994.


Downloaded from by guest on February 13, 2012

Website Builder