Chapter 7.  Factors Related to Professional Practice Indices

This chapter summarizes a series of statistical analyses performed to estimate the extent to which different factors and variables are related to the professional practice indices developed in this study.  It includes the following subsections:

• Introduction
• Factors Related to Professional Practice Indices
• Other Patterns and Relationships
• Conclusions

The professional practice indices presented in the preceding chapters have some limited intrinsic value for policy makers interested in the three professions, but much of the interest by health policy makers in these statistics comes from understanding how the indices are related to the numbers of practitioners in the three professions and, ultimately, to access to patient care, especially for underserved populations.

This chapter examines several hypotheses related to the professional practice indices for NPs, PAs, and CNMs, the numbers of professionals per capita for the three professions in the 50 States, and several measures related to access to and the delivery of care. Given the changes that have taken place in health care and the health workforce in the 1990s, three general patterns were hypothesized with respect to each of the three professions.

• The legal scopes of practice (as measured by the indices described above) increased significantly between 1992 and 2000 across the 50 States, indicating increasing acceptance of the professions by physicians, the public, and government regulators.
• Variations in the professional practice declined between 1992 and 2000, indicating a general convergence or standardization of professional practice environments across the States.
• Positive relationships (i.e., correlations) exist between the professional practice indices and the relative supply of practitioners for the three professions (as measured by practitioner per capita ratios).

In addition to these three key hypotheses, the authors performed supplementary analyses of the relationships between the three professions and physicians (i.e., PAs with all physicians, NPs with all physicians, and CNMs with ob-gyns). Of particular interest is whether or not the three professions and physicians have a complimentary relationship or a substitutive relationship with one another.

Analyses were also performed to assess the extent of relationships between the professional practice indices and other measures of the health care system and the health status of the population. These include HMO penetration in the States and the percentages of States’ populations living in Health Professions Shortage Areas (HPSAs).

The data set compiled for this study includes a number of variables summarized in Table 7-1. The original index for 2000 was developed by identifying specific criteria and weighting schemes that would permit replication of the 1992 indices, and then applying these criteria and weights to conditions in 2000. Because some of the historical files related to the earlier study were not available, it was not possible to identify criteria that permitted replication all of the1992 scores. Several of the criteria used by Sekscenski required application of judgment about points assigned for certain conditions, and the authors were unable to devise a single weighting scheme that would successfully replicate the earlier indices for all the States. Also, the authors did not have complete copies of all statutes and regulations in force in the 50 States in 1992, which complicated the task of assigning scores for specific elements of the indices. Despite these limitations, it was possible to replicate the 1992 scores for 45 of the 50 States.

The data presented in Table 7-2 summarize the original index scores as reported in their NEJM article, along with the results of applying the authors’ best choice of criteria uniformly for all 50 States for 2000. Thus, State scores for 1992 are based on internally consistent criteria and definitions, as are the State scores for 2000. While there is some question about the validity of comparisons of 1992 and 2000 indices, the fact that the authors were able to replicate 90 percent of the State scores of 1992—and that in the cases where replication was not possible the differences were negative—provides a basis for confidence in the comparisons.

The practitioner counts for 1992 were estimated from the article by Sekscenski et al. Counts for later years were obtained from other sources that appear to be the most reliable as indicated in Table 7-1. The data for PAs is believed to be generally accurate and comparable over time. The data for NPs and CNMs, while improving in recent years, have gaps in the early 1990s that will require attention before reliable year-to-year comparisons can be made for this time period.

Variables Used to Test Study Hypotheses

The primary analysis tool used in this study was Spearman’s rank order correlation. This permits comparisons with Sekscenski et al [1994] which also used this technique. The paired t-test was used to compare average values of the Sekscenski indices for 1992 and 2000. In addition, the F-test was used to compare the variances of the Sekscenski indices in 1992 and 2000. In all cases, an alpha level of 0.05 was used to test statistical significance. All tests were performed using SPSS for Windows version 11.0.

Trends in Professional Practice Indices from 1992 to 2000

Table 7-2 summarizes the information in Tables 4-1, 5-1, and 6-1. It shows clearly that on average the 50 States experienced statistically significant increases in the original practice environment indices for all three professions. This is a clear indication that the professional practice options for all three professions expanded between 1992 and 2000.

Table 7-2 also shows that the standard deviation of the original scores across the States was smaller in 2000 than in 1992 for all three professions, and that the difference was statistically significant for NPs and PAs. This is an indication that there has been a general convergence of the professional practice across the 50 States between 1992 and 2000, especially for NPs and PAs.

Table 7-3 presents the three major components (legal status, reimbursement, and prescriptive authority) of the new professional practice indices for NPs, PAs, and CNMs for all 50 States and the District of Columbia for the year 2000. This reveals insights about why one State may have a larger or smaller index for a profession than another State. The table shows that the average overall new index scores for NPs, PAs, and CNMs for 2000 were 74.7, 74.1, and 69.6, respectively, out of a possible total of 100. These scores are significantly lower than the respective original index scores, reflecting the fact that additional options and criteria have been included in the new indices. Readers interested in more detail about the new scope calculations for NPs, PAs, or CNMs may refer to Appendix E, D, or F, respectively.

Comparisons of scores across the three professions, either on average or for individual States are not appropriate. The three indices are based on different criteria and weighting schemes and are not designed to serve as a standard for comparing the professions.

The gaps between the “optimal” scores and the average scores reveal that opportunities for States to increase the index scores for the three professions are generally greatest for prescriptive authority and legal status, and least for reimbursement. The lower a component score for a State below the “optimal”, the greater the opportunity to increase the index through appropriate adjustment in the corresponding criteria.

The standard deviations of the component scores for the new indices show greater variability in scores across the States for prescriptive authority than for legal status and reimbursement. Comparisons of the standard deviations for the components of the three original indices were not made because of difficulties in replicating the 1992 indices for five States.

Numbers of Practitioners

Table 7-4 shows the increases in the numbers of NPs, PAs, and CNMs per 100,000 population    that occurred between 1992 and 2000. Despite some data limitations for the earlier years, the estimates show that the growth has been dramatic, with NPs per capita growing by 190 percent, PAs per capita growing by 70 percent, and CNMs per capita growing by 65 percent over the 8 year period.

Relationships Between the Professional Practice Indices and Numbers of Practitioners

An analysis of the relationship between the three components of each index for the three professions (legal status, prescriptive authority, and reimbursement) across the professions showed positive correlations among the components professional practice indices across States in 2000 (Table 7-5). States with favorable prescriptive authority for PAs also had favorable prescriptive authority for NPs and CNMs. For legal status and reimbursement, NP scores were significantly correlated with CNM scores, while PA scores were not significantly correlated with either NP scores or CNM scores.

Table 7-5

Correlations of Components of the Professional Practice Indices Across the Three Professions
(Coefficients are Spearman rank-order correlations across the 50 states.)

*  = significant at the 0.05 level
** = significant at the 0.01 level

Sekscenski et al found that favorable practice environments, as measured by their practice environment indices, were strongly positively correlated with numbers of the corresponding professionals. This study confirmed this relation for both 1992 and 2000 for all three professions. Table 7-6 shows the Spearman rank order correlations between the 1992 scope indices and 1992 practitioners per 100,000 population, and between the 2000 scope indices and 2000 practitioners per 100,000 population. These correlations confirm that higher professional practice indices are associated with greater numbers of practitioners per capita for all three professions.

Table 7-6 also shows that the professional practice indices are not significantly correlated with the numbers of physicians per 100,000 population for the corresponding years. This is an indication that states with relatively large (or small) numbers of physicians per capita do not have unusually high (or low) professional practice indices.

Table 7-6
Correlations Between Original Professional Practice Indices and Professionals per Capita, 1992 and 2000
(Coefficients are Spearman rank-order correlations across states.)

1992

2000

 ⁺  = Sekscenski index from the original study
 *  = significant at the .05 level
 ** = significant at the .01 level

Relationships Between the Three Professions and Physicians

The nature of the relationship between the three professions and their physician counterparts typically involves some level of dependency on the part of the three professions. PAs work under the supervision of physicians, and most NPs and CNMs work under some formal collaborative or supervisory agreement with physicians. These supervisory and collaborative working relationships suggest a positive correlation between the numbers of physicians and the numbers of the three professions.

There are a variety of factors that influence these relationships, including organizational arrangements, reimbursement policies, historical trends, etc. When the changes in the professions are as dramatic as they have been for NPs, PAs, and CNMs, some of the usual patterns and relationships may be altered. Nevertheless, this preliminary analysis appears to support the presence of a supportive relationship between the three profession and physicians.

If a substitutive relationship existed, one would expect a negative correlation between the physicians per capita and practitioners per capita for that profession, i.e., that States with relatively fewer physicians per capita had relatively more NPs, PAs, or CNMs per capita. Table 7-7 shows no evidence of such a substitution effect. In fact, the data show a statistically significant positive correlation between NPs per capita and physicians per capita, and between CNMs per  capita and Ob-Gyns per  capita in 2000.

Table 7-7

Correlations Between NPs, PAs, and CNMs per Capita and Their Counterpart Physicians per Capita, 2000
(Coefficients are Spearman Rank-Order Correlations)

*  = significant at the .05 level
** = significant at the .01 level

Relationship of Professional Practice Indices to Access to Care

Since one of the stated goals of the programs that originally launched both the PA and NP professions was to increase access to care, it is of interest to assess the extent to which these goals have been achieved. Unfortunately, current national data systems are not able to assign members of the three professions to services provided to underserved populations or to geographic regions identified as shortage areas. The best that can be done at present for all 50 States is to compute correlations between the percentages of  population residing in Health Professions Shortage Areas (HPSAs) and the scope indices and the numbers of practitioners per capita for the respective States.

The strongest correlation with percent of population in HPSAs is physicians per capita. This high negative correlation is expected since a region is designated a HPSA if it has especially low numbers of physicians. It is interesting that CNMs per capita, and not PAs per capita or NPs per capita, is significantly negatively correlated with the percent of population in HPSAs.

It is also interesting that HMO penetration is significantly negatively correlated with percent of population in HPSAs. This suggests that HMOs have a positive impact on access to care, although other interpretations are possible.

Relationship to HMO Penetration

Table 7-9 presents correlations of HMO penetration to the chosen set of variables. Here too the correlations with the scope indices are not statistically significant. The correlations with physicians per capita, NPs per capita, and CNMs per capita are all highly significant, which indicates that HMO penetration is higher in states with larger numbers of these three professions.

* Correlation is significant at the .05 level (2-tailed).
** Correlation is significant at the .01 level (2-tailed).

** Correlation is significant at the .01 level (2-tailed).

General Acceptance of Non-Physician Clinicians.

To get a sense of the extent to which different States have accepted the professions which work closely with physicians, a composite index (equal to the sum of the three new index numbers for 2000) was created. This new index, which is based on all three professions, is not meant to relate to professional practice. It is meant solely to reflect the general acceptance of the professions by government regulators. Oregon had the highest score on this composite index, and South Carolina had the lowest.

This composite index was then translated into a five point scale that rated the general acceptance levels of these non-physician clinicians in the 50 States and the District of Columbia from high acceptance to low acceptance. The results of the translation into the five point scale are displayed in the map in Figure 7-1. The States with the highest general acceptance for the three professions were scattered around the country with higher representation in the Northeast and Northwest, while the lowest general acceptance of the three professions was focused in the Southeast.

[D]

Professional Practice Component Scores

Table 7-10 presents the scores for the three broad components of the new professional practice indices for the three professions in each of  the fifty States and the District of Columbia. Interested readers can use these data to better understand the nature of the practice environments for the three professions in specific States.

Center for Health Workforce Studies, 10/02

Analyses of the 1992 indices provided by Sekscenski et al and the updated indices created by the authors for 2000 indicates that all three professions increased their respective scopes of practice between 1992 and 2000. The increases observed in the professional practice indices for all three professions are generally associated with broader sets of tasks, more autonomous practice environments (i.e., less direct oversight by physicians), and greater opportunities to prescribe controlled substances.

While differences remain in the professional practice index scores across the 50 States, the variation of the index scores declined significantly between 1992 and 2000, suggesting that the 1990s was a period of convergence of professional practice across the 50 States for all three professions. A breakdown of the three components of the 2000 professional practice index demonstrate a convergence in both legal status and prescriptive authority for NPs, PAs, and CNMs across the 50 States. The reimbursement patterns for NPs, PAs, and CNMs converged less across the States than did the other two components of the indices.

Relations With Physicians

In field work conducted in seven States (California, Illinois, New York, North Carolina, Ohio, Oregon, and Texas) in 2001 as part of this study, more than 220 informants (representing the three professions, educators, provider organizations in urban and rural areas, and State and local planners and policy makers) were asked questions about the three professions, including some concerning relations between the medical profession and NPs, PAs, and CNMs. It was interesting that the closer that informants were to actual practice settings in hospitals, clinics, and physician offices, the stronger was the sense that the three professions provide valuable support to physicians as they serve their patients and the public. The idea is that physicians wouldn’t work with NPs, PAs, or CNMs in hospitals, offices, and other settings if they did not believe it was beneficial to their practices and their patients.

Relationship of the Three Professions to Access to Care

One hypothesis of the study, that greater numbers of practitioners in the three professions improved access to health care, especially primary care, could not be tested statistically. Reliable estimates of the numbers of practitioners in the three professions in the 50 States have only recently become available, and reliable estimates of the numbers practicing in shortage areas (i.e., Health Professions Shortage Areas [HPSAs] or Medically Underserved Areas [MUAs], both of which are based on census tracts) or serving underserved population groups are not yet available. Without such information, it is not possible to quantify the extent to which the three professions serve people with low incomes, without health insurance, or with other characteristics associated with lack of adequate health care.

Although it was not possible with the data and other evidence compiled in this study to confirm statistically that a higher professional practice index is related to greater access to health care by undeserved populations, many believe that NPs, PAs, and CNMs “are providing services (especially primary care) to populations that otherwise would be managed by a physician or would not receive services” [Hooker and Berlin, 2002]. Additional information on this provided in Chapter 9.

Further research is warranted on the extent to which greater numbers of practitioners in the three professions improve access to health care, particularly primary care, for underserved populations. Moreover, investigating the relationship between the three professions and HPSAs and MUAs is an important avenue for future research.