Design and sample. The study employed an explanatory, sequential, mixed-method design. Eligible LTC facilities were identified from the publicly available list from the CMS; 57 Medicare/Medicaid-certified skilled nursing facilities/NHs that provided care services to older adults were located within 100 miles of Durham, NC. Of these facilities, 25 responded to a phone call from the researcher requesting participation; 16 declined, 5 responded “maybe,” and 4 agreed; these 4 NHs were designated A through D. The NHs varied by size (from 60 to 125 operational beds) and location (urban/suburban) in North Carolina. The principle investigator contacted the director of nursing (DON) at each facility to explain the study, invite participation, and set a mutually agreeable date for initiating the onsite activities of the 5-day project. Upon agreement, DONs completed a 7-item, facility-related survey before the intervention.
All nursing staff (registered nurses [RNs], licensed practical nurses [LPNs], and certified nursing assistants [CNAs]) employed either part- or full-time at each NH were eligible for study participation. On day 1, nursing staff who were willing to participate and had the ability to read and understand English were given an envelope containing an introductory letter, a consent form, an 11-item demographic and work history survey, and the 19-item nursing culture survey (color coded for RN, LPN, and CNA job categories). No identifiers appeared on the tool other than facility number identifier and the paper color for job category. Participants were asked to review and sign the consent document, complete the NCAT survey, place both items in the envelope provided, and return the sealed envelope to a locked drop box at the nursing station within 3 days. On day 4, the NCAT survey responses were entered into a database, and the mean total score and mean subscale responses were summarized for presentation to nursing staff.
On day 5, all participants were invited to a focus group conducted in each of the 4 facilities to elicit the nursing staff’s perspectives about the findings of the nursing culture assessment in their facility, to explore how well the results captured their nursing culture, and to determine whether and how they perceived nursing culture to be related to implementation of best practices for PrI prevention.
Ethics approval for this study was obtained from the Duke University Institutional Review Board.
Measures. Facility-level data were extracted from public-access Medicare Nursing Home Profiles appropriate to the timeframe of the study.15 Data included the overall star rating (range 1–5) created by the CMS for consumers to use in comparing above and below average quality of care across NHs.6 These data were augmented by the 7 additional facility-related survey items submitted to DONs, which included tenure (in months) in DON position, percent of annual staff turnover, availability of a wound nurse (yes, no), policies on seeking and accepting wound patients, current number of PrIs (yes, no by patient), and current major projects such as new administration or major staff change.
Participating staff-level variables included age, gender, ethnicity, race, level of education, years/months in current position and any other position at current facility, years/months in a similar position at previous facility, employment status (full or part time), and usual working shift.
NCAT administration. The NCAT was used to survey staff perceptions of nursing culture in each facility. The NCAT has been described extensively as a stable, valid, and reliable measure of nursing’s occupational subculture in LTC settings.1,2 Briefly, its 19 declarative statements are scored on a 4-point ordinal scale (total score range 19–76) and converted to normative ranking percentages (0% to 100%)1 associated with 6 factors that include expectations, behavior, satisfaction, teamwork, communication, and professional commitment related to nursing practice.2 Higher scores represent a positively toned nursing occupational subculture with a more beneficial influence expected when implementing a planned change in care.1 Normative rankings are based on data aggregated from 1025 staff members across 54 NHs and represent the percentage of those NHs whose mean score was less than the relevant level.1 Generally, cumulative rank percentiles are referenced in tertiles as high (> 66%), moderate (33%–66%), and low (<33%). For this study, NCAT items were prefaced as follows: “Please think about the practices in place to prevent pressure ulcers in your health care organization. Also, think about your work and professional commitment as you fill out this questionnaire.”
Focus groups. Focus groups were conducted in a private location in each facility during regular working hours by 2 study team members experienced in qualitative research. Facilitators used an interview guide to elicit participants’ perceptions. Each focus group was between 45 and 50 minutes and began with a consent process, followed by a description of the NCAT’s purpose and a presentation of the analysis of the facility’s results. The presentation included a line drawing of a prone human figure with 3 areas of PrI risk highlighted, followed by 7 pie charts, one each for the total NCAT and each subscale normative scores.
Participants reviewed and discussed the results of the NCAT assessment for their facility and their implications for PrI prevention and care. Participants were asked their interpretation of the extent, if any, to which the total NCAT and subscales’ scores represented the culture of nursing related to their facility’s PrI prevention practices, including staff activities, attitudes, and experiences. Sessions were audiorecorded and professionally transcribed verbatim.
Analytic strategy. Frequency and percent of NH profile items, DON items, and raw and standardized NCAT subscale and total scores were analyzed using descriptive univariate analyses. From these data, graphic presentations were prepared for the focus group discussions, including a line drawing of pressure points on a human figure and pie charts representing standardized total NCAT and subscale scores. Quantitative analyses also included analyses of variance with Bonferroni post hoc pairwise comparisons to ascertain whether NCAT total and subscale scores varied significantly by facility.
Focus group transcripts were analyzed using NVivo 9 (QSR International, Melbourne, Australia).16 A systematic process for qualitative data analysis was employed to identify core concepts emerging from the focus groups using the NCAT theoretical framework to guide the qualitative content analysis. The analytic plan derived both a priori codes from the theory and allowed for new themes to emerge post hoc.
Each transcript was considered a unit of analysis, and handwritten notes were made on general themes. The interdisciplinary focus group analysis team included 3 investigators with backgrounds in gerontological nursing, education, and organizational science. To establish consistency among the investigators during coding, all team members coded the same data excerpts from 1 transcript, discussed the coding decisions, and refined code definitions. Next, each coder independently completed coding the remainder of the relevant text of the transcript, and the group of coders met to reinspect all codes. A codebook with definitions was developed to increase trustworthiness. Next, a pair of coders was assigned to each transcript and coded it independently using the established definitions for codes and themes. The coder pairs met to examine the coding that each had applied to the text and compared the coding with each other to establish reliability. Intercoder reliability was greater than 85%. To increase trustworthiness, the full coding team met repeatedly until the full group had reviewed all coded statements for consistency and accuracy of codes. Differences in coding were discussed and resolved by the team. Finally, the PI and 1 other coder met to examine and ensure that all transcripts and coding assignments had been appropriately completed according to correct definitions across group transcripts.
The analysis approach used is well-suited to cross-sectional data17,18 and is particularly useful in health care research because it allows for a priori concepts (eg, salience, sustainability)19 to be combined with inductive analyses. It also creates an explicit audit trail in the data reductions within and between individual facilities and between analytic stages.20,21