Virginia Board of Nursing Workforce Survey

Introduction

Many who might have elected a career in nursing in the past are today choosing other careers. The result is a severe nationwide shortage of nursing personnel, as recently documented by the federal Health Resources and Services Administration. Virginia is no exception to this skilled workforce shortage.
In response, the 2000 Virginia General Assembly amended the code of Virginia by adding a section relating to nursing workforce information. Specifically, this section of the Code reads as follows:

The implementation of the provisions of this act was made by the Virginia Board of Nursing through a contract with the Virginia Tech Center for Survey Research (CSR) to design and conduct a mail survey that would be representative of all types of Virginia resident nurses who are registered, certified, or licensed by the Virginia Board of Nursing. Included are Certified Nurse Aides (CNAs), Licensed Practical Nurses (LPNs), Registered Nurses (RNs), Clinical Nurse Specialists (CNSs), and Licensed Nurse Practitioners (LNPs) -- hereafter referred to in aggregate as "nurses" or "nursing workforce".
These critical data, covering all aspects of the Virginia nursing workforce as specified in the act of the Virginia General Assembly, comprise the first set of current detailed information to begin to better understand the dynamics of the shortages in the nursing workforce in the Commonwealth. This information provides significant detail on the "supply side" of the "supply-demand equation".

Instrument Design

The survey instrument used for this study was developed by the Virginia Tech Center for Survey Research in collaboration with staff members of the Virginia Board of Nursing. Design of the instrument content was focused particularly on those informational components expressly delineated by the amended Code of Virginia, section 54.1-3012.1. Instrument content was also informed by the review of other contemporary surveys of the nursing workforce, and particularly by the survey questions employed in the Nurse Information System (NIS) of the National Council of State Boards of Nursing.
From the outset, the survey instrument was also constructed under the premise that it would be applicable to all types of nurses and nurse aides. Additionally, there were no item content constraints as to work setting specificity. The instrument was also constructed in a manner to be inclusive of those who are not presently in the workforce as well as for those who are presently working outside of the nursing field.
The draft mock-up was then subject to several iterations of changes in item content and wording as recommended by various staff persons at the Virginia Board of Nursing. The revised version of the survey form was then circulated to five RNs, one LNP, and two LPNs for review and additional suggestions. This resulted in the addition of response categories to several items and some additional minor wording changes. This final version was approved by the Executive Director of the Board of Nursing prior to production.
The final approved copy was designed with generous blank space in order to appear easily and quickly answerable. It was printed in a four-page single-fold booklet. It was produced on off-white paper and photo-reduced to a 7" x 9 ½" size in order to enhance the visual appeal and provide a "polished" official presentation to the survey recipient. To further enhance respondent perceptions of importance, the Virginia State seal and the sponsorship by the Department of Health Professions, Board of Nursing, were prominently displayed at the top of the first page of the instrument. Survey forms were consecutively numbered so that returns could be linked back to the mailing list and respondents could be eliminated from receiving subsequent mailings. A copy of the survey instrument text is shown in Appendix A, and an actual survey instrument also accompanies this report.

Sampling Design

The Board of Nursing specified that only in-state residents should be included in the study. After excluding those persons with out-of-state or incomplete residential addresses, the Board of Nursing registry included 116,368 individuals. Each of these persons was assigned by the Board of Nursing to one of five discrete nurse categories based on Board registry information. The number of nurses in the five categories was highly variable: from only 404 CNSs to as many as 62,431 RNs (after removal of RNs who were also in the registry as holding a Virginia license as an LNP or registration as a CNS). Therefore a stratified disproportionate random sampling design was implemented in order to adequately represent each of the five types of nurses with an active license, registration or certification.
Moreover, there was substantial regional variation in the number or resident nurses. Consequently, in order to assure proportional representation by region, the 116,386 nurses were classified into one of six Virginia regions, based on regional definitions used by the Virginia Hospital and Healthcare Association (VHHA). For each type of nurse, the same proportion was sampled for survey inclusion in each of the six Virginia regions.
The actual proportion of registrants sampled varied by nurse group. The sampling proportion within nurse group was determined in a manner so as to provide reasonable assurance that the sampling error associated with responses for each of the five nurse groups would be less than ±5 percent.
Overall, 5.2 percent of all Virginia resident nurses were surveyed. The proportion varied from 100 percent of the 404 holding registration as a CNS, to a random 3 percent of the 62,431 RNs. A total of 6,076 individuals were surveyed. Subsequently, 18 cases were identified as duplicate sample members who held both a license as an LNP and registration as a CNS. These cases were removed from the LNP sample count, and remained in the CNS sample count. Consequently, the final sample size surveyed was 6,058. Full documentation of the sampling design and sample sizes, by type of nurse and by region, is shown in Table 1.

Survey Administration

The research design called for a first wave mailing of the survey to the complete sample, followed by a reminder postcard one week later. A second wave mailing of another survey form was subsequently mailed to non-respondents approximately seven weeks later. In order to distinguish this mail from routine bulk mail, all mailings utilized a first class "live" stamp. Most of the first wave was mailed on Friday, June 8, 2001, with remaining surveys mailed on Monday, June 11. This was followed by a reminder postcard mailed on Monday, June 18. Returned surveys were logged in daily, and a second mailing of the survey form was sent to all first wave non-respondents on Friday, July 27, 2001.

Virginia	CNA Aides		LPN Practical		RN Registered³		CNS Clinical		LNP Practitioner⁴		Total
Region ²	Population (N)	Sample (n)	Population (N)	Sample (n)	Population (N)	Sample (n)	Population (N)	Sample (n)	Population (N)	Sample (n)	Population (N)	Sample (n)
Northern Virginia	3,600	180	2,718	163	16,292	489	101	101	898	297	23,609	1230
Blue Ridge	2,744	137	2,465	148	6,893	207	54	54	355	116	12,511	662
Central	5,830	291	5,420	325	12,654	380	103	103	682	228	24,689	1322
Hampton Roads	7,152	358	6,086	365	14,340	430	91	91	691	232	28,360	1472
Roanoke Area	5,599	280	4,383	263	9,240	277	48	48	357	119	19,627	987
Southwest	2,242	112	2,177	131	3,012	90	7	7	134	45	7,572	385
Virginia Total	27,167	1,358	23,249	1,395	62,431	1,873	404	404	3,117	1,028	116,368	6,058
(Sampling proportion)		(5%)		(6%)		(3%)		(100%)		(33%)		(5.2%)

¹ Counts derived from Virginia Board of Nursing database, as of May 5, 2001. All persons with out-of-state addresses are excluded, as are in-state persons with incomplete residential addresses (4,872 CNAs, 130 LPNs, 332 RNs, 4 CNSs amd 9 LNPs).

²Six regions defined by the Virginia Hospital and Healthcare Association (VHHA) as reported in Indicators of Healthy Communities 2000.

³ Excludes those also holding a Virginia license as LNP or registration as CNS.

⁴ LNPs include those who self-identify as certified registered nurse anesthetist or certified nurse midwife. Eighteen (18) of the sampled LNPs were also included in the CNS sample; these were removed from the LNP population count and sample count, and respondents were included in the data only for CNSs.

Each survey mailing was accompanied by a postage-paid return envelope and a cover letter from the Executive Director of the Virginia Board of Nursing. For the 1,358 CNAs, the salutation was "Dear Virginia Nurse Aide"; with "Dear Virginia Nurse" for all others in the sample. The same salutation variation was retained on the letter accompanying the second wave survey mailing. The text of the cover letter was identical to all. Copies of the first wave cover letter, the reminder postcard, and the second wave cover letter are shown in Appendix B.
All surveys returned from the post office with forwarding addresses were remailed to the new addresses. Additionally, 266 survey forms were returned as undeliverable by the post office and six were returned with the addressee being identified as deceased. After eliminating duplicate returns, a total of 4,405 completed surveys were received. Excluding those that were undeliverable or deceased, the effective response rate was 76.1 percent. The overall response rate ranged from a low of 55.6 percent for Certified Nurse Aides to a high of 86.1 percent of surveys returned by the Clinical Nurse Specialists. The total number of returns for each nurse group, the corresponding response rate, and the associated sampling error for each nurse group are shown in Table 2.
Data entry was undertaken using the Sawtooth Ci3 program which precludes out-of-range entry errors. Once data entry was completed, a simple random sample of 20 percent of survey forms were selected and re-entered by a different data-entry staff person at the CSR. The two independent entries were compared and a keystroke accuracy rate of greater than 99.5

Table 2
2001 Virginia Survey Response Rates and Standard Errors by Type of Nurse

	CNA Aides	LPN Practical	RN Registered	CNS Clinical	LNP Practitioner	Total
Sample Size¹	1,358	1,395	1,873	404	1,028	6,058
Number of Respondents	755	978	1,483	348	841	4,405
Percent Responding	55.6%	70.1%	79.2%	86.1%	81.8%	72.7%²
Sampling Error³	±3.5%	±3.1%	±2.4%	±2.0%	±2.9%	— —

²Subtracting 266 surveys returned from the post office as undeliverable and 6 returned as deceased, the sample size is reduced from 6,058 to 5,786 and the effective response rate is then 76.1%.

percent was ascertained. Any keystroke errors for the quality control random sample were checked against the original survey response form and the error corrected on the data file. All open-ended responses were checked for spelling accuracy and consistency in wording and abbreviations. Those open-ended "other" responses that were obviously similar to a listed coded category were reassigned to that category in the post-processing and editing phase.

Compilation of Respondent Data

Each of the five discrete nurse groups are uniquely different from the others and hence, it is likely unreasonable that responses could be combined across these groups. Indeed, given the different sampling ratios for each group, and the variations in response rates, aggregation of responses from different nurse groups would first require calculation of a multiplier weight to adjust for these differences.
Therefore, for the present report all statistical calculations are made separately for each of the five nurse groups. Compilation of percentage distributions for responses to all close-ended survey questions, for each of the five nurse groups, is shown in Appendix D. Listing of responses to open-ended survey items are presented in Appendix E.

Data Uses and Storage

The data file containing individual responses is provided to the Board of Nursing for possible additional in-house analyses. However, these individual response data will not be released by the CSR to any other entity. This policy is to assure: (1) the guarantee of confidentiality made to individual respondents will not be breached; and (2) the Virginia General Assembly legislative requirement that data shall be provided only in aggregate form and shall not be released in any way which could allow identification of individuals.
This "raw" data file is provided on a diskette which accompanies this report. This is an ASCII data file. The associated codebook is shown in Appendix C as well as included on the diskette. Also included is an SPSS data file with all individual data for close-ended items as well as variable and response category labels and the responses to open-ended items.
The descriptive summary data tabulations, as displayed in Appendix D of this report, are also included in electronic version. This will be transmitted as an electronic file which can be directly loaded onto the Board of Nursing's web site.
The CSR will maintain backup files of the data, the SPSS data file and the program used for tabulating the results for a period of no less than one year. Release of this backup information will only be made to the Virginia Board of Nursing.
The Virginia Tech Center for Survey Research encourages and appreciates acknowledgement of its role in projects. However, in such cases there should be a disclaimer of Center for Survey Research responsibility for interpretations or discussion of implications of results.

VIRGINIA NURSES SURVEY: 2001