Development and Validation of the Perioperative Care Process Adherence Scale

Background and Rationale

One of the key determinants of clinical outcomes in perioperative care is consistent adherence to evidence-based care processes. The Donabedian structure-process-outcome model emphasizes that process quality is closely linked to clinical outcomes, making it critical to render process performance measurable for quality management purposes (Donabedian, 1988). In the perioperative context, this process spans a multi-component chain of care encompassing operating room safety (e.g., surgical safety checklist use), early postoperative monitoring, complication prophylaxis, pain management, Enhanced Recovery After Surgery (ERAS) components such as early mobilization and early feeding, effective communication, patient education, and documentation continuity (Ljungqvist et al., 2017).

The perioperative period constitutes a high-risk, time-critical "acute vulnerability window" in which the surgical stress response, hemodynamic fluctuations, analgesia requirements, early complication development, and safety threats escalate simultaneously within a short time frame. Adherence, therefore, is not merely the existence of a protocol but the timely, correctly sequenced, and well-coordinated implementation of its components. Much of the existing literature on process performance has approached measurement through the lens of "missed nursing care" or care omission, which captures what care was not delivered but does not fully capture holistic adherence to evidence-based care (encompassing presence, timing, continuity, and coordination of care). Existing perioperative nursing scales largely measure patient experience/perception (PREMs) or general care-quality perception, or focus on competency rather than directly modeling process adherence at a performance-near level. This gap motivates the development of a new instrument capable of jointly capturing operating-room and ward-based nursing process adherence, evaluated with rigorous psychometric methods including measurement invariance testing across work settings.

Aims

The purpose of this study is to develop the Perioperative Care Process Adherence Scale (PBSUO) and to evaluate its content validity, construct validity (exploratory and confirmatory factor analysis), convergent and discriminant validity, internal consistency, and temporal reliability in a multicenter sample. Additional aims include testing measurement invariance between operating room and surgical ward nurses and confirming a second-order (hierarchical) factor structure in which sub-dimensions (e.g., operating room safety and team coordination; early monitoring and complication prevention; ERAS adherence; documentation continuity; patient education and communication) load onto a general "overall process adherence" higher-order factor.

Hypotheses

H1 (Construct Validity - Exploratory Phase): PBSUO items will form a theoretically interpretable multidimensional factor structure that explains a significant proportion of common variance (Kaiser-Meyer-Olkin >= .80; Bartlett's test of sphericity significant).

H2 (Construct Validity - Confirmatory and Hierarchical Model): The proposed factor structure will show acceptable to good fit in confirmatory factor analysis (CFI >= .90, preferably >= .95; TLI >= .90, preferably >= .95; RMSEA <= .08, preferably <= .06; SRMR <= .08), and sub-dimensions will load significantly on a single second-order general factor.

H3 (Internal Consistency and Item Performance): The total scale and subscale scores will show adequate internal consistency (Cronbach's alpha >= .70; McDonald's omega >= .70; composite reliability >= .70), and items will show adequate corrected item-total correlations (r >= .30).

H4 (Convergent and Discriminant Validity): Subscales will show Average Variance Extracted (AVE) >= .50 and HTMT ratios < .85-.90, and the total PBSUO score will correlate positively with a measure of perceived care quality/patient safety culture.

H5 (Measurement Invariance): The PBSUO factor structure will be invariant, at least at the metric level and preferably at the scalar level, between operating room and surgical ward nurses (delta CFI <= .010; delta RMSEA <= .015).

H6 (Temporal Reliability): The total PBSUO score will show adequate temporal stability on repeat administration two weeks apart (ICC >= .70).

Study Design and Setting

This is a multicenter methodological scale-development and psychometric-validation study. Data will be collected from nurses actively working in operating rooms and surgical wards at three different hospitals; the multicenter design was chosen to increase sample heterogeneity and strengthen the generalizability of the measurement model.

Scale Development Process

Phase 1 - Item Pool Generation: A comprehensive literature review of behaviors representing evidence-based perioperative process adherence was conducted, drawing on the Donabedian process-quality model, ERAS adherence components, patient-safety indicator literature, and care-process-performance literature. An initial 60-item draft of the PBSUO was developed, structured around six theoretical domains: clinical monitoring and early complication recognition; complication prevention and prophylaxis; ERAS adherence components (early mobilization, early feeding, etc.); patient education and safe communication; documentation and care continuity; and operating room safety and team coordination. All items use a 5-point Likert format (1 = not at all applicable to 5 = always fully applicable), with respondents rating their own clinical practice.

Phase 2 - Content Validity: Content validity will be assessed by a panel of 8-12 subject-matter experts in surgical nursing, operating room nursing, patient safety, quality management, and psychometrics, using a 4-point relevance scale (1 = not relevant to 4 = highly relevant). Items with an Item-Level Content Validity Index (I-CVI) >= .78 will be retained; the Scale-Level CVI (S-CVI/Ave) will be calculated; items deemed not relevant will be revised or removed.

Phase 3 - Cognitive Pilot Testing: The refined draft scale will be piloted with 20-30 operating room and surgical ward nurses to assess comprehensibility, redundancy, completion time, and item clarity, with resulting linguistic refinements incorporated into the final version.

Phase 4 - Large-Sample Psychometric Analysis: The refined scale will be administered to a large multicenter sample for exploratory factor analysis (EFA; principal axis factoring with promax rotation), confirmatory factor analysis (CFA), second-order factor model testing, convergent/discriminant validity (composite reliability, AVE, HTMT), internal consistency (Cronbach's alpha, McDonald's omega), measurement invariance (operating room vs. surgical ward nurses), and temporal reliability (test-retest, ICC).

Participants and Eligibility

Data will be collected from nurses actively working in operating rooms and surgical wards. Inclusion criteria: actively working in an operating room or surgical ward for at least 3 months; directly involved in the clinical perioperative care process; age 18 years or older; voluntary participation with written informed consent. Exclusion criteria: administrative role without direct clinical care delivery; newly oriented staff still in orientation; a communication barrier that would prevent completion of the questionnaire.

Sample Size and Power

The planned sample size is 320-450 nurses. This range reflects the commonly recommended 7-10 participants per item for factor analysis, applied to an anticipated post-CVI/pilot item count of approximately 40-45 items (40 items x 8 = 320; 45 items x 10 = 450). Once the total sample reaches adequate size, the dataset will be split using center- and unit- (operating room/ward) stratified randomization into two independent subsamples: approximately 50% (n ~ 160-225) for EFA and approximately 50% (n ~ 160-225) for CFA and second-order model testing, reducing the risk of overfitting. An a priori power analysis for the convergent-validity correlation (assumed r = .30, alpha = .05, power = .80) indicated a minimum required sample size of approximately n = 84, well below the planned sample. For the measurement invariance analysis, at least n ~ 120-150 is targeted in each of the operating room and ward subgroups. Monte Carlo simulations (1,000-5,000 replications; sample-size scenarios n = 150, 180, 200, 225 for the CFA arm) will be conducted in R to evaluate parameter bias, standard error stability, and convergence rates for the second-order model.

Data Collection

Data collection instruments include the PBSUO itself, a demographic and professional information form, and a brief convergent-validity measure of perceived patient safety culture/care quality. Written informed consent will be obtained from eligible nurses before administration; the questionnaire takes approximately 8-12 minutes to complete. Data will be coded and stored anonymously.

Statistical Analysis Plan

Analyses will be performed using IBM SPSS Statistics and R (lavaan, psych, semTools, and simsem packages), with two-tailed significance set at p < .05. Preprocessing will include missing-data analysis (listwise deletion if missingness <= 5%, otherwise multiple imputation after evaluating the missingness mechanism), outlier screening (univariate |z| > 3.29; multivariate Mahalanobis distance), and evaluation of skewness/kurtosis and floor/ceiling effects (>15%). EFA will use principal axis factoring with promax rotation, with factor retention informed by eigenvalues > 1, scree plot inspection, and parallel analysis, and item retention criteria of factor loading >= .40, communality >= .30, and cross-loading difference >= .20. CFA will use maximum likelihood (or robust ML/WLSMV for ordinal data as needed), with model fit evaluated via chi-square/df < 3, CFI >= .90 (preferably >= .95), TLI >= .90, RMSEA <= .08 (preferably <= .06), and SRMR <= .08. A second-order CFA model will be tested and compared to the first-order model using delta CFI (< .01), AIC, and BIC. Reliability will be assessed via Cronbach's alpha, McDonald's omega, corrected item-total correlations, and composite reliability/AVE. Convergent and discriminant validity will be assessed via AVE, composite reliability, and HTMT, and via the correlation between the PBSUO total score and the convergent-validity measure (Pearson, or Spearman if normality is not met). Measurement invariance (configural, metric, scalar) will be tested via multi-group CFA between operating room and surgical ward nurses, with delta CFI <= .010 and delta RMSEA <= .015 as invariance criteria.

Ethical Considerations

The study will be conducted in accordance with the principles of the Declaration of Helsinki. Written informed consent will be obtained from all participants. Data will be anonymized and used solely for scientific purposes. The study poses no physical risk to participants; some items may cause brief discomfort as they require self-evaluation of professional practice, and participants may decline to answer any question or withdraw at any time without any effect on their employment or professional standing.