Impact of social deprivation and ethnicity on patient-reported outcome measures in primary and revision hip and knee arthroplasty

Abstract

Aims

There is a paucity in the literature on the impact of social deprivation and ethnicity on patient-reported outcome measures (PROMs) in hip and knee arthroplasty. This study aimed to assess the impact of social deprivation and ethnicity on PROMs after hip and knee arthroplasty in a healthcare system which is free at the point of care.

Methods

We reviewed collected data on patient demographic details, comorbidities, and PROMs (EuroQol five-dimension questionnaire, Oxford Hip Score, and Oxford Knee Score), in patients who underwent hip and knee arthroplasty procedures, both primary and revision. The primary outcome was impact of social deprivation on PROMs. Secondary outcomes included the impact of ethnicity, religion, and comorbidities on PROMs.

Results

We identified 10,646 hip arthroplasties and 8,138 knee arthroplasties from the database. Of these, 6,093 patients who underwent hip arthroplasty and 4,681 patients who underwent knee arthroplasty were included. Patients from the most deprived areas, ethnic minorities, and in those for whom English is not their first language had lower preoperative PROMs. Higher levels of social deprivation were associated with higher levels of comorbidity. The trends are similar in revision hip and knee arthroplasty, but the net improvement in the PROMs was significantly lower in patients who underwent revision procedures when compared with a primary procedure.

Conclusion

This study highlights that healthcare inequality exists in the UK among those undergoing hip and knee arthroplasty procedures. This is despite the substantial benefit these procedures can provide, irrespective of social deprivation, ethnicity, or first language.

Cite this article: Bone Jt Open 2026;7(7):956–970.

Introduction

The assessment of patient-reported outcome measures (PROMs) is essential for understanding the impact of interventions on patients.¹ This has been seen and reported in both hip and knee arthroplasty, where functional outcomes such as return to work and pain relief can influence patient satisfaction.² Increasing use of PROMs over the last 20 years has demonstrated the excellent outcomes that can be achieved with joint arthroplasty.² The use of PROMs has also helped to identify important factors that may influence outcomes such as postoperative rehabilitation,³ surgical factors,⁴ lower BMI,⁵ varying rates of PROMs completion,⁶ or a diagnosis of depression.⁵ It has also been suggested that social deprivation and ethnicity are similar such factors.^7,8 While the precise reasons for ethnicity being a factor are poorly understood, the reasons for this are likely to be a complex interaction between factors including financial status, access to healthcare, and occupation. Social deprivation alone has been suggested as an aetiological factor for the development of osteoarthritis.⁹ Social deprivation also influences the interaction between occupation, the inability to be ‘off’ from work, engagement with healthcare, and comorbid disease status. There is a paucity of literature examining the impact of a patient’s ethnicity and social deprivation on PROMs after hip and knee arthroplasty. There is some literature available from developed countries with privately funded healthcare systems that describes how patients from ethnic minorities and those classed as socially deprived have a distrust in the health system. This often leads to a delay in seeking healthcare or accepting treatment.¹⁰ However, with free healthcare that is available through the NHS in the UK, it is not clear whether an individual’s social deprivation status or ethnicity has an impact on outcome from joint arthroplasty surgery in the same way that it may do in healthcare systems that are predominantly funded privately.¹¹ Furthermore, there is a reluctance to proceed with surgery among those from ethnic minorities, as surgery is often considered the last resort and inherently avoided, which often progresses to severe disability.¹² There are some reports that suggest patients from deprived settings have comparable outcomes for knee arthroplasty with those from less deprived backgrounds.^13,14 What is currently unknown is whether it is deprivation status that independently influences the outcome, or whether it is the higher comorbidity rate observed in the socially deprived that primarily affects outcomes in this group of patients.¹⁵

This study aims to assess impact of social deprivation, ethnicity, and comorbidity on PROMs for patients undergoing both primary and revision hip and knee arthroplasty surgery at a single speciality hospital.

Methods

Study approval and design

Approval was obtained from the institutional research department (reference number EL/AG/ROH21ORTH08). This is a retrospective cohort study from a single high-volume centre over a ten-year period.

Data sources

We reviewed the PROMs from patients who had undergone either primary or revision hip and knee arthroplasty from our prospectively collected local database and routinely collected hospital data between October 2012 and September 2022. Data were linked using the hospital identification number. Data collected included age, sex, postcode, comorbidities (using Charlson Comorbidity Index (CCI)),¹⁶ ethnicity, religion, and first language. The PROMs collected were the EuroQol five-dimension three-level questionnaire (EQ-5D-3L) score,¹⁷ Oxford Hip Score (OHS, 2007 version),^18,19 and Oxford Knee Score (OKS, 2007 version).^19,20 PROMs questionnaires were in English. Due to the limitations of the information that had been recorded on the database, we were not able to collect data on BMI, any surgical factors at time of surgery, or the rehabilitation undertaken. We have thus taken a pragmatic approach and have grouped all knee arthroplasties and hip arthroplasties together to overcome this potential confounder, sub-dividing only for whether it was a primary or revision procedure. To make an assessment of social deprivation, the Index of Multiple Deprivation (IMD) was used, which assesses an individual’s social deprivation status based on the postcode.²¹ This is a Department for Communities and Local Government measure and gives information on relative deprivation for small areas.²² It takes into account seven domains: income deprivation, employment deprivation, education skills and training deprivation, health deprivation and disability, crime, barriers to housing and services, and living environment deprivation. This gives a ranking for deprivation across all of the small areas of England, each of which can then be grouped into deciles for analysis.

Baseline demographic characteristics

Table I shows the baseline demographic characteristics of the identified patients.

Table I.

Baseline demographic characteristics of patients included in analysis.

Variable	Hip cohort (n = 6,093)	% of cohort	Knee cohort (n = 4,681)	% of cohort
Age, yrs
< 40	184	3.02	11	0.23
40 to 49	446	7.32	63	1.35
50 to 59	1,026	16.84	621	13.27
60 to 69	1,935	31.76	1,624	34.69
70 to 79	1,884	30.92	1,809	38.65
80 to 89	592	9.72	543	11.60
90 to 99	26	0.43	10	0.21
Sex
Female	3,529	57.92	2,634	56.27
Male	2,564	42.08	2,047	43.73
CCI
0	522	8.57	55	1.17
1	919	15.08	465	9.93
2	1,586	26.03	1,233	26.34
3	1,659	27.23	1,575	33.65
4	860	14.11	888	18.97
5	330	5.42	288	6.15
6	148	2.43	129	2.76
7	51	0.84	44	0.94
8	15	0.25	3	0.06
9	3	0.05	1	0.02
Ethnicity
White British	5,475	89.86	3,896	83.23
White Irish	91	1.49	99	2.11
Any other white background	88	1.44	56	1.20
Asian Chinese	4	0.07	8	0.17
Any other ethnic group	20	0.33	33	0.70
Any other Asian background	16	0.26	60	1.28
Asian British Bangladeshi	0	0.00	6	0.13
Asian British Indian	30	0.49	166	3.55
Asian British Pakistani	16	0.26	99	2.11
Black British African	15	0.25	16	0.34
Any other Black background	17	0.28	19	0.41
Black British Caribbean	59	0.97	61	1.30
Any other mixed background	5	0.08	5	0.11
Mixed white and Asian	6	0.10	4	0.09
Mixed white and Black African	4	0.07	2	0.04
Mixed white and Black Caribbean	19	0.31	14	0.30
Not known	75	1.23	38	0.81
Not stated	153	2.51	99	2.11
Religion
Agnostic	22	0.36	20	0.43
Atheist	67	1.10	40	0.85
Baptist	10	0.16	12	0.26
Celtic Christian	729	11.96	602	12.86
Church of England	799	13.11	706	15.08
Hindu	3	0.05	35	0.75
Jehovah’s Witness	9	0.15	11	0.23
Jewish	8	0.13	4	0.09
Methodist	40	0.66	26	0.56
Muslim	14	0.23	90	1.92
Not religious	48	0.79	49	1.05
Religion not given	3,997	65.60	2,718	58.06
Roman Catholic	297	4.87	286	6.11
Sikh	13	0.21	43	0.92
Other (any other religion with fewer than 10 patients)	37	0.61	39	0.83
Deprivation decile
1	678	11.13	737	15.74
2	615	10.09	553	11.81
3	591	9.70	477	10.19
4	706	11.59	531	11.34
5	776	12.74	579	12.37
6	521	8.55	367	7.84
7	575	9.44	409	8.74
8	523	8.58	295	6.30
9	518	8.50	366	7.82
10	590	9.68	367	7.84
First language
English	5,616	92.17	4,278	91.39
Not specified	459	7.53	318	6.79
Punjabi	3	0.05	36	0.77
Urdu	2	0.03	26	0.56
All other languages	13	0.21	23	0.49
Surgery type
Primary	5,361	87.99	4230	90.37
Revision	732	12.01	451	9.63

CCI, Charlson Comorbidity Index.

Inclusion and exclusion criteria

We collected data from consecutive patients undergoing primary or revision hip and knee arthroplasty at our centre. No patients were excluded at the outset of the study. Patients were only excluded from the analysis if the data on the variables being studied, particularly PROMs or ethnicity data, were not available for analysis.

Outcomes

The primary outcome of interest was the analysis of the association of social deprivation on PROMs in patients who underwent hip and knee arthroplasty procedures. Secondary outcomes were the impact of religion, ethnicity, and comorbidities on the PROMs.

Statistical analysis

Descriptive statistics were used to summarize the demographic characteristics of patients. All statistical analysis was performed using R (v. 4.4.3; R Foundation for Statistical Computing, Austria). Differences between the social deprivation deciles and respective EQ-5D, OHS, and OKS scores were analyzed using both independent-samples t-tests (for direct comparisons between two groups) and Kruskal-Wallis tests to allow for analysis of multiple factors simultaneously. Significance was set at 0.05 for both tests.

Results

From the records reviewed there were data for 10,646 hip arthroplasties and 8,138 knee arthroplasties during the study period. Of the 10,646 hip arthroplasties during this time, there were 9,276 primary hip arthroplasty procedures and 1,370 revision hip arthroplasty procedures. Age ranged from 13 to 95 years (mean 66 years). However, only 6,093 patients in the hip cohort had a complete set of data; 4,270 patients were excluded for incomplete EQ-5D scores, a further 87 were excluded for incomplete OHS, and finally 196 were excluded for incomplete social deprivation data. This is demonstrated in Figure 1. After exclusion there were 16 patients included who were aged below 18 years; we recognize that the OHS has not been validated for paediatric patients. As demonstrated, the majority of patients excluded were of white ethnicity, had no specified religion, and spoke English as their first language.

Fig. 1.

Flow diagram to demonstrate the reasons for exclusions from the hip cohort. EQ-5D, EuroQol five-dimension questionnaire; OHS, Oxford Hip Score.

There were 8,138 knee arthroplasties, with 7,266 primary knee arthroplasty procedures and 872 revision knee arthroplasty procedures. Age ranged from 19 to 93 years of age (mean 69 years). Of the 8,138 patients in the knee cohort, 4,681 patients had a complete set of data; 3,327 were excluded for incomplete EQ-5D scores, a further 122 were excluded for incomplete OKS, and finally eight were excluded for incomplete social deprivation data. This is demonstrated in Figure 2. Again, we did not proportionately lose a large number of patients to follow-up from any ethnic minority, religion, or specific language.

Fig. 2.

Flow diagram to demonstrate the reasons for exclusions from the knee cohort. EQ-5D, EuroQol five-dimension questionnaire; OKS, Oxford Knee Score.

Primary hip arthroplasty

Table II demonstrates the mean preoperative, postoperative, and change in EQ-5D scores for hip patients separated by social deprivation deciles. Social deprivation decile 1 is the most deprived, while decile 10 is the least deprived. Those from the most deprived areas had lower preoperative and postoperative scores compared with those from the least deprived areas. The change in EQ-5D was statistically significantly different between social deprivation decile 1 when compared to decile 10 (p = 0.002, independent-samples t-test), and across all social deciles as a whole (p = 0.029, Kruskal-Wallis test).

Table II.

Mean preoperative and postoperative EuroQol five-dimension (EQ-5D) questionnaire score for hip patients by social deprivation deciles.

Social deprivation decile	Mean preop EQ-5D score	Mean postop EQ-5D score	Change in EQ-5D score
1 (most deprived)	0.29	0.72	0.43
2	0.30	0.73	0.43
3	0.34	0.75	0.41
4	0.38	0.80	0.42
5	0.38	0.79	0.41
6	0.38	0.81	0.43
7	0.41	0.83	0.42
8	0.41	0.83	0.42
9	0.41	0.83	0.42
10 (least deprived)	0.46	0.85	0.39

Figure 3 demonstrates the trend of lower preoperative EQ-5D scores in those from the highest levels of social deprivation.

Fig. 3.

Preoperative EuroQol five-dimension questionnaire (EQ-5D) scores for hip patients grouped by social deprivation decile. The pinch point/solid black bars demonstrate the median, the grey box demonstrates the IQR, the black whiskers demonstrate 1.5 times the IQR, the circles demonstrate values outside of the whiskers of the data, the blue spot demonstrates the mean, and the blue whiskers demonstrate the SD.

The OHS categorized by social deprivation deciles, shown in Table III, demonstrates a similar trend, with those from the highest level of social deprivation having lower preoperative scores and postoperative scores than those from less social deprivation, but with approximately similar overall change in scores following surgery. There is no statistically significant change in the OHS scores between social deprivation deciles 1 and 10 (p = 0.820, independent-samples t-test) or across all deciles as a whole (p = 0.300, Kruskal-Wallis test).

Table III.

Mean preoperative and postoperative Oxford Hip Score (OHS) for hip patients by social deprivation decile.

Social deprivation decile	Mean preop OHS	Mean postop OHS	Change in OHS
1 (most deprived)	16	36	20
2	17	38	21
3	17	38	21
4	19	40	21
5	19	40	21
6	20	40	20
7	20	41	21
8	20	41	21
9	19	41	22
10 (least deprived)	21	42	21

Figure 4 again demonstrates the trend of lower starting OHS in those from the highest levels of social deprivation.

Fig. 4.

Preoperative Oxford Hip Score (OHS) for hip patients grouped by social deprivation decile. The pinch point/solid black bars demonstrate the median, the grey box demonstrates the IQR, the black whiskers demonstrate 1.5 times the IQR, the circles demonstrate values outside of the whiskers of the data, the blue spot demonstrates the mean, and the blue whiskers demonstrate the SD.

Using a Kruskal-Wallis test, the association between the social deprivation deciles and CCI was significant (p < 0.001), confirming that higher levels of social deprivation are associated with higher levels of comorbidity. The changes in EQ-5D (p = 0.015) and OHS (p = 0.007) were also significantly associated with change in CCI, confirming that higher comorbidity leads to worse improvement in PROMs.

Patients who speak English as their first language or are of white ethnicity report a higher preoperative OHS score and EQ-5D than those who speak another language or are from any other ethnicity (Table IV). However, as seen in Table I, patients of white ethnicities and who speak English as their first language make up the majority of the patient cohorts, and therefore identifying statistically significant differences is challenging due to the small numbers of patients in the non-white and non-English speaking cohorts. Interestingly, those patients who do not speak English as a first language and those from non-white ethnicities appear to see greater improvement in OHS and EQ-5D scores than those for whom English is their first language or who are of white ethnicity.

Table IV.

Mean preoperative and postoperative EuroQol five-dimension (EQ-5D) questionnaire scores and Oxford Hip Scores (OHS) for hip patients by first language and ethnicity.

Variable	Mean preop OHS	Mean postop OHS	Change in OHS	Mean preop EQ-5D score	Mean postop EQ-5D score	Change in EQ-5D score
English as first language	19	39	20	0.37	0.79	0.42
Non-English first language	18	42	24	0.36	0.83	0.47
White ethnicity	19	40	21	0.38	0.79	0.41
Non-White ethnicity	17	38	21	0.30	0.77	0.47

Primary knee arthroplasty

When we look at the knee cohort we see similar results, as shown in Tables V and VII below. Figure 5 and Figure 6 demonstrate the trend of lower preoperative PROMs scores for those from the most deprived areas.

Table V.

Mean preoperative and postoperative EuroQol five-dimension questionnaire (EQ-5D) for knee patients by social deprivation deciles.

Social deprivation deciles	Mean preop EQ-5D score	Mean postop EQ-5D score	Change in EQ-5D score
1 (most deprived)	0.29	0.64	0.35
2	0.34	0.66	0.32
3	0.38	0.69	0.31
4	0.38	0.72	0.34
5	0.41	0.71	0.30
6	0.41	0.73	0.32
7	0.41	0.77	0.36
8	0.42	0.75	0.33
9	0.42	0.75	0.33
10 (least deprived)	0.48	0.79	0.31

Table VII.

Mean preoperative and postoperative EuroQol five-dimension questionnaire (EQ-5D) scores and Oxford Knee Scores (OKS) for knee patients by first language and ethnicity.

Variable	Mean preop OKS	Mean postop OKS	Change in OKS	Mean preop EQ-5D score	Mean postop EQ-5D score	Change in EQ-5D score
English as first language	18	35	17	0.38	0.71	0.33
Non-English first language	17	35	18	0.36	0.73	0.37
White ethnicity	19	35	16	0.40	0.72	0.32
Non-white ethnicity	16	32	16	0.29	0.66	0.37

Fig. 5.

Preoperative EuroQol five-dimension questionnaire (EQ-5D) scores for knee patients grouped by social deprivation decile. The pinch point/solid black bars demonstrate the median, the grey box demonstrates the IQR, the black whiskers demonstrate 1.5 times the IQR, the circles demonstrate values outside of the whiskers of the data, the blue spot demonstrates the mean, and the blue whiskers demonstrate the SD.

Fig. 6.

Preoperative Oxford Knee Scores (OKS) for knee patients grouped by social deprivation decile. The pinch point/solid black bars demonstrate the median, the grey box demonstrates the IQR, the black whiskers demonstrate 1.5 times the IQR, the circles demonstrate values outside of the whiskers of the data, the blue spot demonstrates the mean, and the blue whiskers demonstrate the SD.

Table VI.

Mean preoperative and postoperative Oxford Knee Score (OKS) for knee patients by social deprivation decile.

Social deprivation decile	Mean preop OKS	Mean postop OKS	Change in OKS
1 (most deprived)	16	32	16
2	17	33	16
3	18	34	16
4	18	35	17
5	19	35	16
6	19	35	16
7	20	36	16
8	19	36	17
9	19	36	17
10 (least deprived)	21	38	17

OKS, Oxford Knee Score.

Statistical analysis of EQ-5D and social deprivation demonstrates that this change in EQ-5D across the social deprivation deciles is statistically significant (p = 0.009, independent-samples t-test). This reaffirms again that social deprivation does impact on PROMs in knee arthroplasty, specifically the health-related quality of life.

However, when assessing social deprivation deciles against change between mean preoperative and postoperative OKS, the p-value is not significant (p = 0.139, independent-samples t-test), demonstrating that the change in OKS, and therefore specifically knee function, is not related to social deprivation.

A Kruskal-Wallis test demonstrates a significant association between change in EQ-5D and social deprivation decile (p = 0.036), however there is no significant link between social deprivation decile and CCI in this cohort (p = 0.315). It is noteworthy that when assessing for the impact of CCI on change in EQ-5D, there was no significant difference (p = 0.661) in this cohort. Furthermore, there was no significant difference in the change in OKS based on social deprivation decile (p = 0.753) or CCI (p = 0.126).

Revision hip arthroplasty

The mean preoperative EQ-5D and OHS in patients who underwent revision hip arthroplasty were similar to those of the primary hip arthroplasty. This highlights that patients’ quality of life (EQ-5D) and functional scores (OHS) are often similarly low when waiting for a revision as it is while waiting for a primary hip arthroplasty. However, the changes in EQ-5D and OHS are significantly different between primary and revision hip arthroplasty, showing the benefits of a revision hip arthroplasty are lower than after a primary hip arthroplasty (p < 0.001, independent-samples t-test; Table VIII).

Table VIII.

Mean preoperative and postoperative EuroQol five-dimension questionnaire (EQ-5D) scores and Oxford Hip Scores (OHS) for primary hip arthroplasty and revision hip arthroplasty.

Surgery type	Mean preop OHS	Mean postop OHS	Change in OHS	Mean preop EQ-5D score	Mean postop EQ-5D score	Change in EQ-5D score
Primary hip	18	40	22	0.37	0.81	0.44
Revision hip	21	34	13	0.39	0.68	0.29

Similarly to primary hip arthroplasty, in revision hip arthroplasty the change in PROMs from preoperative to postoperative seems to be similar across the social deprivation deciles, as shown in Figure 7 and Figure 8.

Fig. 7.

Mean change (delta) in Euro Qol five-dimension questionnaire (EQ-5D) scores for revision hip arthroplasty across the social deprivation deciles. The pinch point/solid black bars demonstrate the median, the grey box demonstrates the IQR, the black whiskers demonstrate 1.5 times the IQR, the circles demonstrate values outside of the whiskers of the data, the blue spot demonstrates the mean, and the blue whiskers demonstrate the SD.

Fig. 8.

Mean change (delta) in Oxford Hip Score (OHS) for revision hip arthroplasty across the social deprivation deciles. The pinch point/solid black bars demonstrate the median, the grey box demonstrates the IQR, the black whiskers demonstrate 1.5 times the IQR, the circles demonstrate values outside of the whiskers of the data, the blue spot demonstrates the mean, and the blue whiskers demonstrate the SD.

Revision knee arthroplasty

The mean preoperative EQ-5D scores and OKS were lower in patients who underwent revision knee arthroplasty compared with those who underwent primary knee arthroplasty. This demonstrates that patients awaiting a knee revision arthroplasty may often be worse than they were before their primary joint arthroplasty. We also see lower mean postoperative scores for revisions than we do for primaries, again demonstrating that revision knee arthroplasty does provide improvements, but does not achieve the same benefit seen in primary joint arthroplasty. The changes in EQ-5D and in OKS were significantly different between primary and revision knee arthroplasty (p < 0.001, independent-samples t-test). This is demonstrated in Table IX.

Table IX.

Mean preoperative and postoperative EuroQol five-dimension questionnaire (EQ-5D) and Oxford Knee Score (OKS) for primary knee arthroplasty and revision knee arthroplasty.

Surgery type	Mean preop OKS	Mean postop OKS	Change in OKS	Mean preop EQ-5D score	Mean postop EQ-5D score	Change in EQ-5D score
Primary knee	19	35	16	0.39	0.73	0.34
Revision knee	15	27	12	0.27	0.53	0.26

Revision knee arthroplasties also achieved similar change in the PROMs across the social deprivation deciles as shown in Figure 9 and Figure 10.

Fig. 9.

Mean change (delta) in EuroQol five-dimension questionnaire (EQ-5D) scores for revision knee arthroplasty across the social deprivation deciles. The pinch point/solid black bars demonstrate the median, the grey box demonstrates the IQR, the black whiskers demonstrate 1.5 times the IQR, the circles demonstrate values outside of the whiskers of the data, the blue spot demonstrates the mean, and the blue whiskers demonstrate the SD.

Fig. 10.

Mean change (delta) in Oxford Knee Score (OKS) for revision knee arthroplasty across the social deprivation deciles.

Discussion

This paper highlights that in the UK population, despite free access to healthcare, disparities in preoperative and postoperative PROMs exist in patients undergoing hip and knee arthroplasty procedures. In patients from the most deprived areas, ethnic minorities and in those who English is not their first language, we see lower preoperative EQ-5D scores, OHS, and OKS. The improvement in PROMs after hip and knee arthroplasty in our results are significantly above the minimal clinically important difference (MCID) quoted in other studies^23,24 and highlights that hip and knee arthroplasty still achieves excellent improvements in PROMs. However, our study also shows that while those from the most deprived areas see a statistically significant difference in their change in EQ-5D compared with those from the least deprived areas, this is still not enough to ‘catch-up’ and leaves them with lower final scores.

There are published reports suggesting that worst outcomes are seen in patients from higher levels of social deprivation.^11,25-31 However, many of these studies are based on privately funded healthcare systems, where social deprivation creates a significant barrier to early access to healthcare. In the UK, this barrier should not exist as the NHS provides care which is free at the point of access. This study therefore confirms that despite the lack of this potential barrier, social deprivation still leads to health inequalities^32,33 suggesting other confounding factors which influence the poorer outcomes are seen in patients from socially deprived backgrounds.

A potential confounding factor is the increased rates of comorbidities⁷ seen in patients from higher levels of social deprivation. Our study has shown that there is a statistically significant association between increasing Charlson Comorbidity Index and higher levels of social deprivation in the hip cohort. The Charlson Comorbidity Index was also seen to have a tiered impact on the PROMs with a higher comorbidity index leading to lower PROMs. However, we did not see the same association between social deprivation decile and Charlson Comorbidity Index in the knee cohort. This may be due to the differences within the demographic details of the groups, however it is difficult to know the definite reason. Regardless, our data show that social deprivation and comorbidities are closely related factors but both appear to influence PROMs.

Furthermore, there may be problems with the use of PROMs in patients from the most deprived areas. Social deprivation has been shown to impact on completion rates of PROMs, therefore leading to a lack of data to analyze.³⁴ This potentially introduces a bias on the reliability of the data for assessment in this group.

Patients from high levels of social deprivation are more likely to have lower preoperative scores,^{11,26,27,29,35} which was a feature clearly demonstrated in our results. Lower preoperative scores are likely to be multifactorial in nature but include difficulties such as poor awareness or understanding of healthcare processes, difficulty accessing healthcare, and delaying treatment due to financial pressures and the need to continue working.

Lower preoperative scores do not necessarily predict outcome³⁶ and instead change in PROMs has been suggested as the best way to measure outcomes in those from deprived backgrounds.²⁷ While the primary hip and knee cohorts did see a statistically significant difference in the change in EQ-5D scores between the most deprived and least deprived areas, the mean numbers are very close and this significance is likely influenced by the size of the cohort and do not represent a clinical difference. Most notably, those from the most deprived areas start from a lower preoperative EQ-5D score and are only likely to achieve a proportionate change, hence a lower postoperative EQ-5D. We did not recalculate a MCID for our cohorts as we did not have a reliable measure to allow an anchor-based method of calculation.

A similar pattern of worse outcomes is seen in those from ethnic minorities.³⁵ The poor outcomes are further compounded when we study those from ethnic minorities from deprived backgrounds.³⁷ Our results demonstrate that those from non-white ethnicities and those who spoke a first language other than English had lower preoperative scores. This is likely to be multifactorial but once again highlights that a cohort of patients, who should receive the same standard of care, appear to access healthcare later with worse PROMs. While the OHS and OKS are validated in several languages, they are only used in English at our institute. However, our results also show the excellent outcomes which can be achieved in this cohort. The change in EQ-5D was higher for both non-white ethnicities and in whom English was not their first language, for both hip and knee arthroplasty, when compared with white ethnicities and in those who English was their first language. This was not tested for significance due to the small cohort size but provides reassurance about the improvement in quality of life seen in these cohorts.

Our study has several limitations. Firstly it is a cohort study and only allows for retrospective analysis of the prospectively collected data. It inherently suffers from the drawbacks of analyzing the data retrospectively and may be a potential source of bias. As highlighted in the discussion, PROMs have been reported to have worse completion rates in those from higher levels of deprivation. Our results also suffer from loss to follow-up which could potentially be a further source of bias. As shown in Figure 1 and Figure 2, as well as Table I, there does not appear to have been a significant loss to follow-up from any one social deprivation decile, ethnicity, or religion. There is evidence of impact of physiotherapy on PROMs. Details of therapy input was not assessed as a part of this study but is likely to influence the outcome. OHS and OKS are validated in several languages but during the study period were only used in English. Patients who do not speak English are likely to rely on family and friends to complete the score for them which may not be reliable.

In conclusion, this study highlights that there exists healthcare inequality in the UK among those undergoing hip and knee arthroplasty procedures despite the substantial benefit these procedures can provide to patients irrespective of social deprivation, ethnicity, or first language. Facilitating easier access to orthopaedic and physiotherapy services with outreach into areas of social deprivation, and into ethnic minority communities may help facilitate earlier intervention and better outcome. Improved communication with information in languages commonly used in the community will also enhance engagement. Ultimately, if we can identify and treat patients earlier, and importantly before their PROMs drop below the levels seen in patients from less social deprivation, we may be able to ensure that all patients achieve the same excellent PROMs provided by arthroplasty surgery.

Take home message

- This study highlights that healthcare inequality exists in the UK among those undergoing hip and knee arthroplasty procedures.

- This is despite the substantial benefit these procedures can provide, irrespective of social deprivation, ethnicity, or first language.

Data Availability

The datasets generated and analyzed in the current study are not publicly available due to data protection regulations. Access to data is limited to the researchers who have obtained permission for data processing. Further inquiries can be made to the corresponding author.