Introduction:

3.2 Clinical characteristics according to OA status

3.3 Anatomical distribution of documented OA

3.4 WOMAC and quality-of-life measures

3.5 Radiographic findings

Radiographic knee assessment was available for 175 participants. K/L grade ≥2 was present in 72/175 (41.1%). This is a radiographic proportion among the subgroup with available knee radiographs and should not be interpreted as the prevalence of OA in the complete 500-participant dataset.

The multivariable model included all 500 participants and 177 OA events. Model discrimination was modest (AUC 0.660). VIFs ranged from 1.01 to 1.19, indicating no important multicollinearity.

3.7 Missing data and follow-up variables

Missingness was concentrated in WOMAC, radiographic and follow-up variables. WOMAC variables were available for 150/500 participants, knee K/L grade for 175/500, follow-up pain and follow-up WOMAC for 135/500, and new joint involvement for 441/500. Because follow-up ascertainment was incomplete and the primary objective was characterization of documented OA, follow-up variables were not included in the primary analysis.

4. Discussion: In this retrospective study of 500 adults aged ≥50 years attending a service hospital, OA was documented in 35.4% of records. The knee was the predominant primary affected site, accounting for 67.2% of documented OA cases. Participants with documented OA were older, had a modestly higher mean BMI, reported greater pain and morning stiffness, and had lower EQ-5D-5L index values than those without documented OA.

The 35.4% figure should not be interpreted as population prevalence. The study was facility-based, retrospective and dependent on clinician documentation rather than standardized prospective case ascertainment. Indian systematic reviews have demonstrated substantial heterogeneity in knee-OA estimates according to population, setting and diagnostic criteria.[2–4] The present findings are therefore most appropriately interpreted as a description of OA documented in this service-hospital dataset.

The predominance of knee involvement is consistent with established OA epidemiology.[1,2] Increasing age and higher BMI were independently associated with documented OA in the adjusted analysis. Previous joint injury showed the largest adjusted association among the categorical predictors, with more than twice the odds of documented OA. This is compatible with the recognized relationship between prior joint injury and subsequent degenerative joint change, although the retrospective design and record-based exposure ascertainment do not establish temporality for every observation.

Low physical activity was independently associated with documented OA, whereas sedentary hours per day were not. This finding requires cautious interpretation because reduced activity may be a consequence of pain and functional limitation rather than an antecedent exposure. Thus, the observed association should not be interpreted as evidence that low physical activity causes OA. The distinction also illustrates that categorical physical-activity level and reported sedentary time may capture different aspects of behaviour.

Participants with documented OA had substantially lower EQ-5D-5L index values than those without documented OA. The Indian EQ-5D-5L value set provides an established framework for deriving preference-based health-state values in India.[8] However, the present analysis is descriptive and does not establish population utility norms or treatment-related changes in health-related quality of life.

WOMAC findings were exploratory because data were available in only 30% of participants and were markedly unbalanced between groups. The absence of statistically significant differences should therefore not be interpreted as evidence of equivalence between participants with and without documented OA. The small non-OA subgroup substantially limits precision.

Among participants with available knee radiographs, 41.1% had K/L grade ≥2. This estimate cannot be directly compared with the 35.4% proportion of documented clinical OA because the radiographic subgroup was selected by availability of imaging and the primary outcome included OA at multiple anatomical sites. Interpretation of K/L grading also depends on consistent application of the classification system.[6]

The study has potential clinical relevance for service-hospital practice because the recorded burden of symptoms and reduced quality of life indicates that OA may represent a meaningful source of morbidity among older adults attending such facilities. Current international recommendations emphasize individualized, multicomponent non-pharmacological management, including education, exercise, weight management and support for function and participation.[9,10] The present study did not evaluate treatment effectiveness and therefore cannot determine whether such interventions improve outcomes in this population.

5. Strengths and limitations

5.1 Strengths

• A relatively large single-centre dataset of 500 adults aged ≥50 years.

• Availability of demographic, anthropometric, behavioural, comorbidity, symptom and health-related quality-of-life variables.

• A prespecified multivariable model that excluded variables forming part of the clinical diagnostic pathway, reducing the risk of incorporation bias.

• Availability of radiographic K/L grading in a defined subgroup.

• Explicit reporting of missingness and avoidance of unsupported reconstruction of screening, exclusion or missing-data counts.

5.2 Limitations

• The retrospective record-based design permits information and selection bias and limits causal inference.

• Facility-based sampling limits generalizability to the wider Indian population.

• The primary outcome was record-documented OA rather than a prospectively standardized diagnosis.

• Radiographs were available for only 175 participants, creating potential selection bias in radiographic analyses.

• WOMAC was available for only 150 participants, with only 16 participants without documented OA represented in that subset.

• The source or measurement instrument used to categorize physical activity was not documented in the supplied materials.

• Only final K/L grades were available, preventing calculation of inter-observer agreement.

• Incomplete follow-up data prevented robust longitudinal analysis.

• Residual confounding from unmeasured factors cannot be excluded.

6. Conclusion: Among 500 adults aged ≥50 years in a service-hospital retrospective dataset, OA was documented in 35.4% of records and most commonly involved the knee. Older age, higher BMI, previous joint injury and low physical activity were independently associated with documented OA. Documented OA was also associated with greater pain and morning stiffness and lower health-related quality of life. Because the study was facility-based and retrospective, the observed proportion should not be interpreted as population prevalence and the observed associations should not be considered causal. Prospective, multicentre Indian studies using standardized diagnostic criteria and more complete functional and radiographic assessment are warranted.

7. Declarations

Ethics approval and consent to participate

The study was approved by the institutional ethics committee. The requirement for individual informed consent was waived because this was a retrospective, minimal-risk study using de-identified clinical records.

Consent for publication

Not applicable; no individually identifiable patient information is reported.

Availability of data and materials

The underlying patient-level dataset is subject to institutional confidentiality restrictions. A suitably de-identified dataset may be made available from the corresponding author or institution, subject to applicable institutional approvals.

Funding: No specific funding was received for this study.

Authors' contributions: SS, AK and SD contributed to study conception, protocol development, data acquisition, data analysis and interpretation. SS, AK, SD, AR and AC contributed to manuscript drafting and final approval of the submitted version. All authors approved the final manuscript.

Acknowledgements: None.

Declaration of conflicting interests: The authors declare no conflicting interests.