Annals of Allied Health Sciences

Why ANCOVA Should Be Standard Practice in Rehabilitation Clinical Trials?

Hazrat Bilal, Aamna Bibi — Wed, 31 Dec 2025 00:00:00 +0000

Randomized Controlled Trials (RCTs) in rehabilitation often involve heterogeneous patient populations and small sample sizes, making precise estimation of treatment effects challenging. Analysis of covariance (ANCOVA) is gaining prominence as a superior statistical method in RCTs, especially in rehabilitation sciences, due to its ability to adjust for baseline differences and increase statistical power compared to unadjusted or change-score analyses.^1,2 This adjustment is important in rehabilitation trials where there is a huge variation in functional status at the baseline level.

Recent methodological studies reveal that ANCOVA provides unbiased, more precise and highly statistically enhanced estimation of treatment effects, even when model assumptions are moderately violated.³ For example, Wang et al. reported that ANCOVA ensures the integrity of inferential conclusions and enhances power in trials with continuous and binary outcomes. ⁴ Additionally, ANCOVA can decrease the required sample sizes by up to 25%, maximizing resource efficiency in clinical research.⁵

In rehabilitation sciences, where intervention protocols are complex and multi-component and the outcomes are multi-dimensional, ANCOVA’s ability to control for baseline heterogeneity enhances interpretability and trial rigour.⁶ Therefore, ANCOVA should be the standard analytic approach in rehabilitation RCTs to ensure accurate and efficient evaluation of therapeutic interventions.

Before using the ANCOVA, it must be noted that the dependent variable must be continuous and must be normally distributed, the covariate must be a continuous quantitative variable, the levels of the qualitative variable must be independent ⁷, the dependent variable and covariate should have a linear relationship⁸ the positive or negative sign and magnitude of the correlation coefficient at each level of the qualitative variable should be similar between the dependent variable and the covariate. ⁹ In other words, there should be equality of regression slopes⁷ there should be no relationship between the independent variable and the covariate variable. ¹⁰

How to correctly report the ANCOVA results:

Revealing the correlation coefficient and statistically significant P-value of assessing the association between the covariate and dependent variable. ¹⁰
Revealing the insignificant relationship between the independent variable and covariate variable, and thus the homogeneity of slope of regression lines.
Tabular representation of means of the dependent variable before and after the adjustment of the impact of the covariate, as well as revealing the p-value of the means comparison separately.

References

Silva CN, et al. The effectiveness of progressions of difficulty during an exercise program in older individuals: a randomised clinical trial. Arch Gerontol Geriatr. 2025; 108:104975.
Johnson L, et al. Statistical and methodological considerations for randomised controlled trials in physical medicine and rehabilitation. PM&R. 2023;15(3):234-245.
Vickers AJ, Altman DG. Empirical comparison of baseline covariate adjustment methods in clinical trials. Trials. 2022;23(1):112.
Wang R, et al. Analysis of covariance (ANCOVA) in randomised trials: robustness and precision gain. Biometrics. 2023;79(2):345-356.
Yang S, Tsiatis AA. Robustness of ANCOVA in randomised trials under model misspecification. Biometrics. 2023;79(1):123-134.
Silva CN, et al. Application of ANCOVA in rehabilitation randomised clinical trials: a systematic review. J Rehabil Med. 2024;56(1):45-52
Field A. Discovering statistics using IBM SPSS Statistics. Sage Publications Limited; 2024 Feb 22.
Rutherford A. ANOVA and ANCOVA: a GLM approach. John Wiley & Sons; 2011 Oct 25.
Karpen SC. Misuses of regression and ANCOVA in educational research. American journal of pharmaceutical education. 2017 Oct 1;81(8).
Schneider BA, Avivi-Reich M, Mozuraitis M. A cautionary note on the use of the Analysis of Covariance (ANCOVA) in classification designs with and without within-subject factors. Frontiers in psychology. 2015 Apr 21; 6:474.

Association of balance and levels of functional disability among children with autism spectrum disorder

Wed, 31 Dec 2025 00:00:00 +0000

Background: In recent years, there has been an increased number of diagnosed cases of autism spectrum disorder (ASD), a persistent developmental condition resulting from both genetic and environmental influences. Beyond challenges in cognition, sensory processing and social interaction, children with ASD often exhibit physical impairments including reduced muscle strength, unsteadiness and balance difficulty. This study was designed to determine the association of balance and levels of functional disability among children with ASD.

Methodology: Analytical cross-sectional study was conducted in Lahore at Lahore Garrison Institute of Special Education and Help autism Pakistan. Children aged 5-12 years with mild to moderate autistic symptoms were included in study while children with any comorbidities and severe autistic conditions were excluded. Balance was assessed using Pediatric Balance Scale and functional disability levels were assessed using World Health Organization Disability Assessment Schedule 2 Children and Youth scale.

Results: Out of 377 children, 285 were males and 92 were females. Among children with low balance (n=340), 78.8% had severe disability, while 21.2% exhibited a moderate disability. In children with near to normal balance (n=37), 8.1% have no disability while 89.2% have mild disability and 2.7% have severe disability. Statistical analysis result showed a significant relationship (P < 0.001).

Conclusion: There is association between balance and levels of functional disability which indicates that autistic children with lower level of balance have higher levels of disability and those with near to normal balance have reduced levels of functional disability.

The association of prolonged sitting with musculoskeletal discomfort among office workers of Peshawar, Pakistan: A Cross-sectional study

Shakir Ullah, Yasir Ashraf, Mumtaz Ali, Amjad Ali — Wed, 31 Dec 2025 00:00:00 +0000

Background: Excessive sitting behavior is a potential hazard for a large number of adverse health outcomes. The increasing trend toward computer-based work has led to substantial sedentary time in workplace settings, especially within the office environment. Prolonged sitting at work is recognized to have a detrimental impact on health in general. However, its precise repercussions on the musculoskeletal system are unclear. This study sought to determine the prevalence of musculoskeletal discomfort in this population and to find its relationship with prolonged sitting.

Methodology: A cross-sectional design was employed, enrolling n = 200 office workers, aged 18-60 years, and with at least a year of experience using a convenience sampling. Data collection utilized the Nordic Musculoskeletal Questionnaire, and Pain intensity was assessed using the NPRS. Associations among variables were analyzed using the chi-square test.

Results: The mean age of participants was 34.83 ± 8.22 years, and they spent an average of 6.10 ± 1.39 hours per day seated. Overall, 88% of people reported having musculoskeletal discomfort, with the lower back being the most common site (64%). Extended sitting duration was significantly associated with musculoskeletal discomfort in the cervical (P = 0.041), elbows (P = 0.045), and wrist/hands (P = 0.002). Additionally, a significant association emerged between musculoskeletal discomfort and both gender (P = 0.023) & exercise routine (P < 0.001).

Conclusion: Prolonged sitting adversely impacts the musculoskeletal system among office workers. An active working environment & short breaks during working hours are therefore recommended to reduce sitting time and ultimately the health hazard.

Evaluation of cardiorespiratory fitness among undergraduates of Rehman Medical Institute: a cross-sectional study

Wed, 31 Dec 2025 00:00:00 +0000

Background: Cardiorespiratory fitness is the capability of the circulatory and respiratory system to supply oxygen to the body. It is the interplay between the cardiac and pulmonary systems to ensure an effective supply of oxygen to the body. In the young generation, cardiorespiratory fitness predicts several health conditions, such as cardiometabolic health and mental health. This study was designed to determine the cardiorespiratory fitness among the undergraduate students of Rehman Medical Institute.

Methods: This research study was cross-sectional in which 322 students were recruited from five colleges of Rehman Medical Institute (RMC, RCRS, RCD, RCN, RCAHS), and systematic sampling was used. Data was gathered using a self-administered questionnaire and a 3-minute step test. The analysis was done using SPSS version 22.

Results: Out of a total population of 322, the maximum number of participants, 41.6% (n=134), were in excellent fitness levels, and the least number of participants were in the poor fitness category with a percentage of 1.9% (n=6). Among the five colleges of RMI, Rehman College of Allied Health Sciences (RCAHS) ranked first with the highest excellent fitness score of 38.8%, while Rehman College of Rehabilitation Sciences (RCRS) ranked last with an excellent fitness score of 9.7%.

Conclusion: This thorough study brought us a set of knowledge regarding the cardiorespiratory fitness of students. From the present study, it is concluded that the majority of undergraduate students had excellent fitness scores, with a percentage of 41.6% and Rehman College of Allied Health Sciences had the highest excellent fitness score, which is 38.8%.

Frequency of lumbar spondylosis in patients with lumbar herniated discs on magnetic resonance imaging

Danial Hanan, Rimsha Badar, Mariam Tariq, Urwah Iftikhar — Wed, 31 Dec 2025 00:00:00 +0000

Background: Lumbar Spondylosis is a degenerative condition characterized by disc wear, disc dehydration, osteophyte formation, disc herniation, spinal stenosis, spondylolisthesis and ligament hypertrophy. This study was designed to determine the frequency of lumbar spondylosis in patients with lumbar herniated discs on Magnetic Resonance Imaging.

Methodology: It was a retrospective study conducted from August 2024 to December 2024 at Chughtai Lab department of Radiology in Lahore, Pakistan. A total of 45 patients with backache were evaluated on Siemens Avanto 1.5T MRI machine. All patients underwent MRI Lumbosacral Spine examination. Non-probability convenient sampling technique was used and data was analyzed using SPSS version 26. Bar charts and Histogram were used to display the data. A p-value less than 5% was taken for statistical significance

Results: Out of 45 patients, 17 (37.7%) were males and 28 (62.2%) were females. Mean age was 53.6214.615 years. Different types of discs herniation were identified including disc bulges in 30 (66.6%), disc protrusion in 11 (24.4%), disc extrusion in 3 (6.6%) and disc sequestration in 2 (4.4%) patients. Only 15 (33.3%) patients out of 45 were diagnosed with lumbar spondylosis, while it was absent in 30 (66.6%) patients

Conclusion: Lumbar spondylosis is a rare diagnosis, presenting with several patterns of disk degeneration especially discs herniation and MRI is the gold standard modality for its accurate analysis.

Comparing the efficacy of virtual reality versus traditional model-based learning methods of anatomy; a perspective of physical therapy students

Madiha Anees, Bilal Khan, Mariyam Kifayat — Wed, 31 Dec 2025 00:00:00 +0000

Background: Anatomy is a major subject taught in early foundation years of medical education which provide information regarding the basic structure of body and their spatial relationship. Initially students learnt anatomy by the visual images on textbooks, atlas, cadavers and 3D models of body structures in labs. However, as the digital world is advancing, and the virtual reality has created an environment of computer-based learning which has created multidimensional model for learning anatomy through immersive or semi-immersive methods. This study was designed to compare the efficacy of virtual reality versus traditional model-based learning methods of anatomy among the physiotherapy students.

Methodology: This was a cross-sectional study conducted over a period of 1 month on undergraduate physical therapy students at Rehman College of Rehabilitation Sciences. 142 students of 1^st, 2^nd and 3^rd year were included in the study through universal sampling. After the approval from the principal of the college the participants were approached, and their informed consent was taken. The 23-item questionnaire was distributed among the participants. Data was analyzed using SPSS version 22. Percentages and frequencies were used to present demographics and graphically presented by bar chart. To find association between independent and dependent variables, Chi square test was applied.

Results: The response rate was 87.3 % out of which 72.6% were females while 24.7% were males having mean age ± 20.2 ± 1.37 and majorly 38.1% students were from 1^st year. On average 34.43% and 39.82% students agree regarding the perception of learning through virtual method and traditional method of anatomy respectively. Between the group there was no significance regarding which method of learning was better for students (p value = 0.5) and no association was found with independent variables (age, gender, and year of study).

Conclusion: The use of VR is as effective as traditional model method of leaning anatomy among physical therapy students however, professionally trained faculty could play a better role in understanding of anatomy learning.