This was an observational cross-sectional study. Each participant was assessed once for preoperative anxiety prior to surgery, and once for POP (12 h – 24 h after surgery). The inclusion criteria were adult patients (≥ 18 years) with an American Society of Anaesthesiologists (ASA) physical status of I–III who were scheduled for elective orthopaedic or general surgery between 07:00 and 16:00. Patients were excluded if they were unable to provide informed consent for the study; had pre-existing anxiety or chronic pain disorders; required postoperative high care or intensive care unit (ICU) admission; underwent day case surgery; and did not speak, read or understand English. The data collection sheets were in English to ensure consistency in administering the questionnaire, as the APAIS has not been validated in local South African languages. Using the tool in its original language reduced the risk of altering its meaning.

This study was conducted at Helen Joseph Hospital (HJH), a 630 bed tertiary hospital in Johannesburg, South Africa,²⁰ affiliated with the University of the Witwatersrand. The hospital provides a range of specialist services, excluding paediatrics, obstetrics, and gynaecology. Recruitment and data collection were conducted from 01 July 2024 to 07 September 2024.

Participants were recruited by the researcher through convenience sampling. All eligible patients present in the ward during the principal researcher’s visits (07:00–16:00) were approached consecutively; those unavailable at these times were not included, as the investigator had other clinical duties and was not ward-based full-time. Of the estimated 213 patients presenting for elective surgery during the study period (112 orthopaedic and 101 general surgery), 130 were enrolled, and 105 (49.2%) were ultimately included in the analysis.

Enrolment occurred in the ward either the day before or on the morning of surgery, after preoperative counselling by the surgeon and anaesthetist. This timing was chosen to capture anxiety that is most likely to influence intraoperative and postoperative outcomes. Data collection involved structured patient interviews supplemented by a review of patient records and anaesthetic charts. Interviews were conducted at two time points: preoperatively in the ward (to assess anxiety) and 12 h – 24 h postoperatively (to assess pain). During the postoperative interview, patients were asked to recall and report when they had experienced their peak pain. Patients were provided questionnaires, which the researcher also read aloud to ensure comprehension and accommodate limited literacy. At the time of preoperative interviews, no enrolled patient had documented anxiolytic premedication.

A structured data sheet was used, comprising a preoperative section (demographics, comorbidities, surgical history, and APAIS questionnaire) and a postoperative section (NPRS scores, peak pain context, and number of pain assessments with assessor identity).

Postoperative pain was assessed using the NPRS, explained verbally to patients and with a visual aid on the data sheet. Patients were asked: ‘On a scale from 0 to 10, where 0 means “no pain” and 10 means “the worst pain you can imagine”, please rate your pain intensity’. Patients rated their pain by stating or pointing to a whole number, which the investigator recorded. Pain was categorised as no pain (0), mild (1–3), moderate (4–6) or severe (7–10). Each patient reported: (1) current pain at the time of interview, (2) the highest pain recalled since surgery, and (3) the context of that peak pain: during movement (including coughing if immobile); upon awakening; after spinal anaesthesia wore off, or at the time of the interview.

These points were selected as they represent common periods of pain fluctuation in the immediate postoperative phase and are easily recalled. If a patient reported an NPRS ≥ 7 as their current pain, the investigator immediately informed the managing team after the interview, who attended to the patient according to ward protocols.

Preoperative anxiety was assessed using the APAIS, developed by Moerman et al.¹⁹ It is a six-item questionnaire rated on a five-point Likert scale (1 = not at all, 5 = extremely), comprising four items on anxiety (two related to anaesthesia and two regarding surgery) and two items on the need for information. The anxiety subscale ranges from 4 to 20, with scores ≥ 11 indicating clinically significant anxiety. The APAIS information subscale ranges from 2 to 10, with scores 2–4 indicating low need, 5–7 average need, and 8–10 high need for information.¹⁹

At the time of protocol development, theatre statistics indicated an average of N = 102 elective general and orthopaedic procedures per month. Based on previous studies,¹² the anticipated prevalence of moderate-to-severe POP was p = 0.62.¹² The required sample size for estimating a single proportion with a 5% margin of error at the 95% confidence level was calculated using Cochran’s formula (Equation 1): n0=z2*p*(1−p)e2[Eqn 1]

Where:

This yielded n₀ = 362. After applying the finite population correction for the monthly sampling frame (N = 102), the adjusted minimum was 81. Allowing for a 15% margin for missing data and non-response, the final required sample size was n = 96, which was rounded up to a target of 100 participants. No separate a priori powering was performed for the secondary objective of examining associations between preoperative anxiety and POP; as such, the study should be regarded as providing pilot data for this question, providing effect estimates to inform future fully powered studies.

Preoperative and postoperative survey data were captured in REDCap by the principal investigator and analysed using R software. Descriptive statistics summarised patient demographics, clinical indications, surgical categories, preoperative anxiety, and POP scores. Categorical variables were presented as frequencies and percentages. Continuous variables were reported as means with standard deviations (s.d.) when approximately normally distributed, otherwise as medians with interquartile ranges (IQR: 25th–75th percentiles). Group comparisons for continuous outcomes used independent-samples t-tests when normality was reasonably satisfied; otherwise, the Mann–Whitney U test was applied. For paired comparisons, we used the paired t-test when the distribution of paired differences met normality assumptions; otherwise, the Wilcoxon signed-rank test was used. Correlations were assessed using Spearman’s rank correlation (ρ), and 95% confidence intervals (CIs) were obtained via nonparametric bootstrap (2000 resamples; percentile method; fixed seed). Associations between categorical variables were assessed using Pearson’s χ² test (Fisher’s exact test when expected cell counts were < 5). For effect-size estimation, univariable logistic regression models yielded crude odds ratios (CORs) with 95% Wald CI. To evaluate the association between preoperative anxiety and POP, we fitted a multivariable logistic regression model to adjust for confounding; results are presented as adjusted odds ratios (aORs) with 95% CI. Model diagnostics for inference assessed sparse cells/separation, multicollinearity (VIFs), influential observations (deviance residuals, leverage, Cook’s distance), and events-per-parameter. All tests were two-sided with α = 0.05.

Ethical clearance to conduct this study was obtained from the Human Research Ethics Committee (Medical) of the University of the Witwatersrand (M240119 MED23-10-356). Permission was obtained from the Research Committee of Helen Joseph Hospital (HJH). Written informed consent was obtained from all participants with the understanding that they could withdraw from the study at any time.

A total of 130 patients were initially enrolled, representing 61.0% of surgical patients booked for elective surgery during the study period. Twenty-five were excluded overall: 6 did not meet the inclusion criteria and 19 had their procedures postponed and therefore did not undergo surgery. Consequently, 105 participants completed both the preoperative questionnaire and the POP assessment; only these 105 were included in the final analyses.

Baseline demographic and clinical characteristics of the study population are presented in Table 1. Of the 105 participants, 42 (40.0%) were female and 63 (60.0%) were male. The overall mean age was 45.5 ± 14.7 years; by sex, the mean age was 46.0 ± 12.4 years in females and 45.2 ± 16.2 years in males. The most common age band was 41–60 years. Using Welch’s two-sample t-test, mean age did not differ by sex (95% CI: −4.71 to 6.37; p = 0.77).

Fifty-nine participants (56.2%) were scheduled for orthopaedic surgery and 46 (43.8%) for general surgery. Sixty-two (59.1%) had undergone at least one prior surgical procedure, and 40 (38.1%) reported one or more comorbidities, most frequently human immunodeficiency virus (HIV) infection and hypertension. Taken together, the typical participant was male, aged 41–60 years, without comorbidities, with prior surgical experience, and scheduled for orthopaedic surgery.

Figure 1 illustrates the types of procedures performed in orthopaedic and general surgery. The most common orthopaedic procedures involved open reduction and internal fixation (ORIF) of the lower limbs, as well as ORIF of the upper limbs, while hernia repairs and open abdominal surgeries were most frequently performed in general surgery.

Preoperative anxiety was present in 26 out of 105 (24.8%) patients. Patients with a high need for information showed a strong positive association with preoperative anxiety (crude OR = 8.43, 95% CI: 3.08–23.04; χ² = 20.114, degree of freedom [df] = 1; p < 0.001). Female sex was associated with higher odds of preoperative anxiety compared with male sex (crude OR = 2.63, 95% CI: 1.06–6.50; χ² = 4.507, df = 1; p = 0.034). There was no evidence of association for type of surgery, previous surgical experience, comorbidities or age group (all χ² p > 0.20). These findings are summarised in Table 2.

No association was observed between time to interview and POP, whether treated continuously (Spearman ρ = −0.017, p = 0.862) or in bins (12 h – 18 h: n = 33/48 [68.8%] vs 19 h – 24 h: n = 36/57 [63.2%]; χ²(1) = 0.362, p = 0.548). Consequently, time to interview was omitted from the primary analyses.

At the time of the interview (12 h – 24 h postoperatively), 69 out of 105 patients (65.7%) had moderate-to-severe POP. In addition, 84 out of 105 (80.0%) reported that their peak POP occurred between the end of surgery and the time of interview. These results are presented in Table 3.

Orthopaedic surgery was associated with higher odds of moderate-to-severe POP compared with general surgery (crude OR = 5.19, 95% CI: 2.17–12.46; χ² = 14.624, df = 1; p < 0.001) as presented in Table 4. There was no evidence of association between moderate-to-severe POP and sex, age group, anaesthesia administered (general vs neuraxial), prior surgical experience, high need for information or number of pain assessments (all p > 0.19).

In this study, 65.7% of participants reported moderate-to-severe POP at 12 h – 24 h after surgery. Pain scores did not vary by interview timing, suggesting consistently inadequate analgesia rather than measurement bias. Persistent pain across time points is concerning and highlights systemic gaps in pain management.

These findings align with reports from other African studies, which show a prevalence of POP varying from 70–75.^8,9 This alignment likely reflects shared systemic challenges, which include limited resources, inconsistent pain assessment, and the absence of structured acute pain protocols.⁷

Globally, compared with high-income countries, our results reflect a significantly higher burden. For example, a Dutch study of 1490 patients reported 41% moderate-to-severe pain on the day of surgery and 30% on the first postoperative day.²² The lower prevalence can be linked to availability of adequate resources allowing greater use of multimodal analgesia and structured acute pain services. A South African audit in Cape Town demonstrated that strengthening acute pain services reduced mean visual analogue scale scores, underscoring the potential impact of dedicated protocols.¹¹

Several factors may explain the high prevalence in our study. Although not formally assessed, it was observed during the study that the hospital lacked a formal acute pain service, and advanced analgesic techniques such as epidurals, wound catheters, and patient-controlled analgesia (PCA) were not used. Only 17% of patients received peripheral nerve blocks, while PCA pumps were unavailable and epidurals were limited by the hospital’s monitoring requirements. Analgesic stockouts and reliance on surgical teams for prescriptions – who may have limited expertise in pain management – further compounded the problem.

Pain peaks were commonly reported on ward arrival or after neuraxial block resolution consistent with a study by Murray et al.¹² These peaks suggest suboptimal pain management during critical transition periods. Peak pain on ward arrival may reflect inadequate intraoperative analgesia, while peaks after block resolution may indicate delayed reporting once motor function returned. Timely reporting of motor function return signals waning neuraxial anaesthesia and enables pre-emptive intervention before severe pain develops. Delays highlight gaps in patient education on self-reporting. As Gao et al. note, effective self-reporting is essential for pain management.⁷ Strengthening perioperative communication and targeted education for patients and staff may improve pain control during these vulnerable phases.

Orthopaedic patients experienced significantly higher rates of POP compared with general surgery patients (OR = 5.19, 95% CI: 2.17–12.46). Despite more structured assessments in orthopaedic wards by nursing staff using the Visual Analogue Scale (VAS), outcomes remained poor (as suggested by the lack of association between number of assessments done and the prevalence of POP), suggesting that assessment alone is insufficient without timely intervention. Informal observations also suggest high pain scores may result from subtherapeutic dosing, cautious prescribing due to fear of side effects, standardised regimens, and delays in adjusting analgesic doses until ward rounds. Together, these findings emphasise the need for comprehensive acute pain pathways supported by training and adequate resources.

In this study, 24.8% of patients experienced preoperative anxiety, lower than both the global pooled estimate of 48% and the African estimate of 56% reported in a meta-analysis by Abate et al.²¹ A local study by Lamacraft et al. found a prevalence of 48% among elective orthopaedic patients,¹⁶ notably higher than our rate of 25.4% (in orthopaedic subgroup analysis). Although their study shares similarities with ours, the difference may be explained by its timing during the coronavirus disease 2019 (COVID-19) pandemic, when baseline anxiety was elevated. Furthermore, our participants had already received preoperative counselling from both the anaesthetist and surgeon before completing the questionnaire. However, despite receiving prior counselling, some patients still reported heightened preoperative anxiety, as high need for information was strongly associated with higher levels of preoperative anxiety, suggesting that the counselling may not have fully addressed their concerns.

Female patients were found to be 2.6 times more likely to report anxiety, consistent with global literature.^21,23 In contrast, previous surgical experience, comorbidities, age group, and type of surgery showed no significant association with anxiety levels in our cohort, despite being identified in prior literature as potential determinants of preoperative anxiety. Interestingly, although international data suggest previous surgery reduces anxiety,²¹ our study and a local study by Lamacraft et al.¹⁶ found no such effect, pointing to the possible influence of contextual or cultural factors

Anxiety about the surgical procedure was significantly greater than anxiety about anaesthesia, indicating that surgical concerns predominate. This highlights the need for counselling to more directly address surgical fears, while also improving awareness of anaesthetic risks, which may be less tangible to patients.

Neither Spearman correlation nor chi-square testing demonstrated a significant association between preoperative anxiety and POP. The crude odds ratio (95% CI: 0.59–4.19) had a wide interval, suggesting that results might differ with a larger, adequately powered sample. Adjusting for sex, age group, and type of surgery as potential confounders did not alter the findings. Interestingly, when focusing on procedure-related anxiety, a significant but weak monotonic correlation with POP was observed (ρ = 0.21, p = 0.036).

Literature on this relationship shows mixed findings. While some studies have reported positive associations,^24,25 a recent meta-analysis by Shebl et al.²⁶ found no such link. Regardless, the impact of preoperative anxiety on the healthcare system remains significant. If an association exists, addressing anxiety could improve pain management, reduce costs, and provide clinicians with an additional tool in a multimodal strategy. Conversely, if no association exists, preoperative anxiety still adversely affects other surgical outcomes, while POP continues to demand resources, thereby compounding the burden on the healthcare system.

Interpretation of this relationship in the South African context can be complex. Inadequate analgesia may overshadow the influence of anxiety on POP, while broader social, cultural, and economic influences may also play a role, as highlighted by Yadeta et al.²⁷ Thus, adequately powered studies employing multivariable models are essential to elucidate these relationships and guide targeted interventions.

This study has several limitations. It was conducted at a single tertiary hospital, which may limit generalisability; however, as part of the teaching hospitals affiliated to the University of the Witwatersrand, it provides a useful baseline for future work. Given the variability in POP management across institutions, targeted studies within specific hospitals or units remain important to identify context-specific challenges and inform tailored interventions.

The relatively small sample size (N = 105) restricted statistical power, particularly for assessing associations between preoperative anxiety and POP; these findings should therefore be viewed as exploratory. Anxiety was measured using the APAIS, which has not been validated in South African languages, and only English-speaking patients were included. This may have introduced bias, as vernacular languages often carry more nuanced meanings that could improve comprehension and response accuracy.

Postoperative pain was assessed once, within a 12 h – 24 h window, rather than at standardised time points, which may not fully capture the pain trajectory. Finally, perioperative analgesic practices – including prescribing patterns, multimodal strategies, and intraoperative techniques – were not systematically assessed. These provider-related factors likely influenced pain outcomes and limit interpretation of the independent role of preoperative anxiety.