Estimation of Expenditure and Challenges Related to Rehabilitation After Knee Arthroplasty: A Hospital-Based Cross-Sectional Study

Abstract

Background

Expenditure for rehabilitation following knee arthroplasty for osteoarthritis- and rehabilitation-related challenges following discharge to home after surgery is not available in the Indian context.

Objectives

To estimate cost of rehabilitation and document challenges in following rehabilitation advices, from a patient perspective.

Methods

We conducted a hospital-based cross-sectional study of patients visiting the orthopedic department at a tertiary care public-funded hospital in New Delhi and included those who recently (less than 4 months) underwent primary knee arthroplasty for osteoarthritis. A trained physiotherapist not involved in clinical care collected information on expenditures incurred after discharge from hospital, patient’s ability to recall the advices given by the physician and challenges they experienced using a semi-structured questionnaire. We report median costs by category of direct and indirect cost and used linear regression to explore determinants of cost.

Results

We interviewed 82 consecutive patients (mean age 60.8 years and 68% females) with median time since surgery of 28 days. More than half (52%) sought some support for physiotherapy. The median cost of rehabilitation was INR 18,395 (Interquartile-range 11,325–27,775). Direct medical cost contributed to 74% of total cost (32% fee for services, 21% medications and lab investigations, 21% assistive devices). Twenty percent higher costs were incurred among those undergoing bilateral knee surgery after adjusting for age, sex, income, and type of physiotherapy support sought. Challenges were related to recall of advices, not understanding the recovery process and pain management.

Conclusion

About half patients undergoing knee arthroplasty seek support for rehabilitation after discharge to home contributing to a major portion of expenses incurred during rehabilitation. Cost-effective support mechanism for home-based rehabilitation is required for improving patient rehabilitation experiences.

Supplementary Information

The online version contains supplementary material available at 10.1007/s43465-021-00405-6.

Background

Physical rehabilitation is a crucial component for acute and post-discharge recovery following knee arthroplasty [1]. The rate and extent of functional ability at 6-month post-surgery could be a function of adherence to the rehabilitation protocols [2, 3]. Post-surgery, individuals face several challenges in adhering to rehabilitation due to pain during exercise, low self-esteem, lack of motivation, and lack of social support [4]. To address these challenges, several rehabilitation strategies that are clinic or hospital-based have been tested against home-based strategies [1, 5]. Current evidence suggests, no added benefit with non-home-based therapies. However, the directness of evidence [1, 5] is limited as these studies have been done in the context of developed countries and no data exist from developing countries.

The standardized care pathways following discharge are varied across countries [6–9] and health systems within a country [8]. The trend for arthroplasty is increasing in developing countries and in India. However, there is lack of information on post-surgery rehabilitation challenges [10]. In India, home-based rehabilitation is the norm [7] and it is unclear how many seek professional help for rehabilitation following discharge. While knee arthroplasty is a cost-effective procedure for severe arthritis [11], the cost of rehabilitation in US was found to be almost half of cost of the surgery [12] after primary total knee arthroplasty. In India, the government reimbursement package for knee arthroplasty was capped at INR 80,000 under the National Health Insurance Scheme (Ayushman Bharat) [13] and this does not include the post-discharge rehabilitation expenditures. Little information is known on the expenditures incurred by the patients during the initial months following surgery.

Hence, the primary objective of this cross-sectional study was to estimate the expenditures incurred by the patients for rehabilitation. Secondary objective was to document the challenges faced by individuals during rehabilitation period.

Methods

We conducted a cross-sectional study at orthopedics outpatient department (OPD) of All India Institute of Medical Sciences, a tertiary care referral hospital at New Delhi between December 2019 and March 2020. We approached consecutive patients visiting the physiotherapy OPD who were operated by any of the orthopedic surgeon of the department. We included patients who underwent unilateral/bilateral primary knee replacement (uni-compartmental/total) for degenerative arthritis within 4 months from the date of interview. We excluded those who were operated for sports injury, trauma, malignancy, inflammatory arthritis, infection or if their duration of surgery extended beyond 4 months from date of interview. The routine rehabilitation protocol at this hospital after surgery is summarized in Appendix 1. At the time of discharge, the exercises to be followed at home are verbally instructed to the patient by the physiotherapist in the presence of an accompanier. A printed pamphlet illustrating the key exercises is also given for reference.

Formal sample size calculations were not performed, but we planned to collect data from at least 100 individuals. Following written consent, information on socio-demography, clinical features, surgical details and analgesic intake, self-reported knee pain and function (0 was worse and 100 was best state), range of motion, rehabilitation expenditure, and challenges related to rehabilitation were collected using paper-based pre-tested questionnaire by a trained physiotherapist not involved in the clinical care (Appendix 2).

We measured active flexion and extension of both knees in supine position using a digital goniometer (Therapy Plus TM Digital Goniometer). We categorized the operated knees into two groups, having flexion below 90° and above 90°.

We administered a single item questionnaire, lower extremity activity scale (LEAS) [14] that represented their self-perceived activity level in a 18-point scale after translating to Hindi. (one indicating confinement to bed and 18 equated to ability to perform vigorous activity daily). Knee Injury and Osteoarthritis Score JR [15] was translated to Hindi.

We collected rehabilitation adherence using a 17-item tool in Hindi (Appendix 2) designed specifically for this patient population adopted from a questionnaire reported by Shin and colleauges [16]. This pre-piloted questionnaire consisted of items relating to family support, focus on recovery, and scheduling ability and pain. Participants were asked to choose one of the four responses ranging from “Strongly disagree to Strongly agree”. The total score was an unweighted sum of responses rescaled to 0–100 with higher scores indicating better adherence. The last item was an 11-point Likert scale to rate self-adherence to home-based exercises between 0 and 10 (0 was “Not at all doing exercises” and 10 indicated “Completely doing exercises as instructed)”.

We calculated total expenditures and out-of-pocket expenditure (OOPE) incurred between the time since discharge after surgery and time of interview. The total expenditure included direct medical (clinician/physiotherapist fee, medications and assistive devices, lab investigations and any procedures), direct non-medical expenses (travel, food and accommodation) and indirect expenses due to wages lost by the patient and the accompanier. OOPE was total expenditure minus amount reimbursed by health insurance.

We asked participants to recall advices given to them by surgeon and physiotherapist at the time of discharge without any prompt. If they could not recall then they were prompted. Responses to each of the 10 items were grouped into two categories “those who could recall the advice with or without prompt” and “those who could not recall the advices with prompt”. This included cold therapy, strengthening exercises, mobilization, pain management, assistive devices, calcium and Vitamin-D, and avoidance of low sitting, squatting and cross sitting, and prevention from falls. We reported an overall response that categorized respondents to those who could recall correct responses for all the 10 items and who could not.

Finally, two open-ended questions were asked to know patients impression on challenges due to pain and difficulties in understanding the instructions given at the time of discharge.

Statistical Analysis

We described the sample population and the key measures (cost and challenges) using summary statistics. We reported overall expenditure, direct, and indirect expenditure separately. The dependency of overall expenditure with time since surgery was assessed by Spearman’s correlation coefficient. The determinants of total expenditure were explored using multiple linear regression adjusted for age in years, sex, monthly income, laterality of surgery, and self-physiotherapy (Yes/No). The total expenditure (dependent variable) was log transformed to normalize the distribution. All analyses were performed using Stata 15 and beta coefficients are reported along with 95% CI. The association of total expenditure with rehabilitation adherence score and ability to recall advices was also analyzed using simple linear regression.

Results

Of the 93 patients approached, 82 were eligible and interviewed. The median duration since surgery was 28 days (IQR 17.5–62). Mean age was 60 years and majority (68%) were females. Seventy percent were from urban area and half received less than high school education. Of the total, 64 had TKR and 44% had a bilateral procedure (Table 1).

Table 1.

Socio-demographic and clinical characteristics of patients at 4 months of surgery

Participant characteristics	(n = 72)
Mean age, years (SD)	60.8 (6.9)
Female, n (%)	49 (68.0)
Urban locality, n (%)	50 (69.4)
Delhi/NCR	23 (31.9)
Married and living with spouse, n (%)	68 (94.4)
Extended family type, n (%)	36 (50.0)
Above high schooling	37 (51.3)
Household monthly income > INR 30,000	37 (51.3)
Health insurance scheme, n (%)	13 (18.0)
No comorbidities, n (%)	28 (38.8)
Laterality of surgery, n (%)
Left	16 (22.2)
Right	24 (33.3)
Bilateral (L)	32 (44.4)
Type of surgery, n (%)
UKR	8 (11.1)
TKR	64 (88.8)

SD standard deviation, NCR national capital territory, UKR uni-compartmental knee replacement, TKR total knee replacement

At time of interview, joint pain score was 65 (100 is no pain), LEAS was 8.7 (which indicates on an average they could go in and outside house without assistance). The mean flexion ROM of the operated knee was 90.4° and 48% had flexion less than 90° at the time of interview.

Almost half (48.6%) undertook exercises at home without assistance. Of those who sought assistance (n = 37), 80.7% sought assistance by trained physiotherapist at home (Table 2). The mean perceived rehabilitation adherence score was 79.1 (100 meaning best adherence). However, only 26% could recall all the instructions provided to them at the time of discharge.

Table 2.

Knee pain and function, physiotherapy and adherence related information

Information	(n = 72)
Median number of days since surgery (IQR)	28 (17.5–62)
Joint pain scorea (SD)	65.1 (16.5)
Current analgesic intake, n (%)
Related to operated knee, n (%)	53 (73.6)
Related to other reasons, n (%)	2 (2.7)
Mean score of LEAS (SD)	8.7 (2.3)
Mean flexion ROM of operated knee (SD) (n = 104 knees)	90.4 (16.9)
Flexion below 90°, n (%) (n = 104 knees)	54 (51.9)
Place of availing physiotherapy, n (%)
Assistance at home (n = 31)	31 (43.0)
By skilled Physiotherapist (n = 25)	25 (80.7)
By untrained attendant (n = 6)	6 (19.4)
At hospital and clinic (n = 6)	6 (8.3)
Self with no assistance (n = 35)	35 (48.6)
Mean rehabilitation adherence (RAQ)b score (SD)	79.2 (11.6)
Mean global rehabilitation adherence rating, 0–10 (SD)	8.2 (1.3)
Ability to recall all the instructions with/without prompt, n (%)	19 (26.3)

ROM (range of motion) = range of flexion-range of extension lag

^aScore of 0–100, 100 indicate perfect knee health and 0 indicates total knee disability LEAS: Lower extremity activity scale, 18-level scale with 1 representing confinement to bed & 18 represents ability to do vigorous activity daily

^bRAQ: Lower score means low adherence to the treatment (0–100 range of possible score)

The total median expenditure incurred since time of surgery and time of interview was 18,395 INR (IQR 11,325–27,775) and the OOPE was almost the same (Table 3). Even though, 18% were covered by health insurance these expenses were not covered. The break-up of money spent by categories (direct and indirect) are illustrated in Fig. 1. The fee for services contributed to the largest proportion of the total cost (32%) followed by assistive devices (21%) and medications (16%). In the non-medical category, travel expenses contributed to most of the expenses.

Table 3.

Median total expenses by its components up to 4 months of surgery

Expenditure	Median expenses (INR), (IQR), (range) (n = 72)
Total expenditure	18,395 (11,325–27,775), (3100–102,500)
Total direct expenses	18,080 (10,150–25,025), (3100–98,499)
Total Indirect expenses	0 (0–1350), (0–25,000)
Total direct medical	11,600 (6960–20,500), (1000–90,500)
Fee for service	0 (0–9800), (0–63,000)
Medications	2933 (2000–4000), (0–20,000)
Assistive devices	4400 (3400–5800), (0–18,850)
Lab investigations	300 (0–1200), (0–5000)
Hospital procedures	0 (0), (0–8200)
Total direct non-medical	2250 (775–6000), (0–33,320)
Travel	1550 (600–4000), (0–20,000)
Accommodation	0 (0), (0–22,000)
Food	0 (0–200), (0–11,000)
Total indirect expenses	0 (0–1350), (0–25,000)
Productivity loss (patient)	0 (0), (0–9000)
Productivity loss (accompanier)	0 (0–1000), (0–25,000)
Total out of pocket expenditure (OOPE)	18,055 (11,025–29,775), (300–102,500)

Fig. 1.

Break-up of costs for rehabilitation (patient perspective) (n = 72)

The total expenditure was not correlated with time since surgery (r = 0.19). In the unadjusted analyses, Higher total expenditure was noted among female patients, those undergoing bilateral knee surgery and among those who sought professional help for physiotherapy (Table 4). After adjusting for age, sex, and monthly household income, the expenditure was 20% higher (95% CI 4%, 36%) among those who underwent bilateral knee surgery and 62% lower (95% CI − 96%, − 27%) among those that did not seek professional help for physiotherapy. There was no difference in joint score (p = 0.16), ROM (p = 0.27) and LEAS (p = 0.42) between those who sought and did not seek professional physiotherapist support.

Table 4.

Determinants of total expenditure (n = 72)

Determinants	Unadjusted beta coefficient (95% CI)	p-value	Adjusted beta coefficient (95% CI)	p-value
Age in years	− 2% (− 4%, 0%)	0.09	− 1% (− 3%, 0.9%)	0.26
Sex
Male	Ref 39% (2%, 76%)	0.04	Ref 9% (− 27%, 45%)	0.62
Female
Residence
Rural	Ref 23% (− 15%, 61%)	0.24	–	–
Urban
Schooling
Below high school	Ref 9% (− 27%, 44%)	0.63	–	–
Above high school
Income
Below INR 30,000	Ref 25% (− 10%, 60%)	0.15	Ref − 0.3% (− 34%, 34%)	0.99
Above INR 30,000
Family type
Nuclear	Ref 5% (− 31%, 41%)	0.78	–	–
Extended
Type of surgery
UKR	Ref 5% (− 51%, 62%)	0.85	–	–
TKR
Laterality of surgery
Unilateral	Ref 25% (7%, 41%)	0.006	20% (4%, 36%)	0.014
Bilateral
Physiotherapy
By non-self	Ref − 70.5% (− 99%, − 39%)	< 0.001	− 62% (− 96%, − 27%)	0.001
By Self
Rehabilitation adherence questionnaire score	0.8% (− 0.7%, 2.3%)	0.29	–	–

Recall of instructions provided at the time of discharge to home was poor. Response to individual items are shown in Fig. 2. Most (n = 31) felt instructions were not clear to them without being specific of what was not clear, and others specifically reported lack of clarity in duration and frequency of exercise, duration of use of assistive devices and a lack of clarity about recovery and return to daily routines (Fig. 2). Recall of advices was not associated with total expenditure incurred (p = 0.34). Of the total, 88% felt pain following surgery as a major challenge to deal with (Fig. 2). They reported disturbed sleep due to pain or difficulty in change in posture during sleep. Back pain was a major issue (n = 12) and some (n = 13) had pain specifically during walking/bending /stair climbing only.

Fig. 2.

Recall of advices (10 items) and challenges related to rehabilitation instructions and pain after discharge based on responses of an open-ended question

Discussion

This hospital-based cross-sectional study among individuals who underwent knee replacement for end stage arthritis from India aimed to provide expenditure estimates incurred in the early post-discharge period and document the challenges faced during rehabilitation. The mean total expenditure was close to one-fourth (INR 18,395) of the total cost of TKR (INR 80,000 as per government capping for TKR). Direct medical expenditure contributed the most (74%) and fee for services at home or a clinic contributed to 32% of total. Patient found it difficult to comprehend the physiotherapy instructions and felt managing pain in the initial weeks of surgery as the main challenge. Self-reported adherence to rehabilitation was high however only one-fourth could recall all the instructions given at the time of discharge. This study highlights the need for additional support to undertake prescribed physical therapy and pain management after discharge to home.

The key strengths of this study was the novelty of the information related to post-discharge rehabilitation collected in the context to Indian settings. There are certain limitations in this study. First, though this study was conducted at a large referral hospital that caters to a very large population, these findings are limited to those undergoing knee arthroplasty from a public-funded hospital. Second, a pre hoc sample size calculation was not possible due to lack of information from prior studies. The imprecision around the effect estimates of determinants of expenditure is a limitation. However, the descriptive data presented in this study will help in sample size estimations for future studies in this topic. Finally, the questionnaires administered in Hindi will require further psychometric analyses in the Indian context. Hence, the findings of the rehabilitation adherence score should be interpreted with caution and could be another limitation of this present study.

Existing evidence with respect to cost of various strategies of post-knee arthroplasty rehabilitation to improve adherence are mainly from developed countries. A study from USA, using the Medicare database (1998 to 2009) reported an increasing trend to discharge to home with home health services instead of nursing or a rehabilitation unit. They also demonstrated that irrespective of the discharge status, the cost incurred on physiotherapy was substantial [17]. A cost-comparison study from Turkey demonstrated the total cost of home-based self-exercise without support was half (38 USD) compared to a clinic-based physiotherapist supported program (65 USD) with no difference in functional status at 2 years [18]. A meta-analysis of four trials [19] concluded the cost of hospital-based rehabilitation was higher (SMD 0.42 (95% CI 0.04, 0.80) compared to unsupported home-based program. However, another meta-analysis of three trials comparing home-based support program with an in-hospital program did not show a difference in cost at 3 months (Mean difference 9.54 USD; 95% CI − 215.03 to 234.11) mainly because of the cost of home-visits by physiotherapist [20]. A study from Australia demonstrated no advantage in in-hospital rehabilitation program compared to discharge to non-hospital-based programs with respect joint specific and health scores and provider costs were substantially higher with in-hospital rehabilitation [21].

To summarize, accumulating evidence suggest in-hospital [21] or clinic-based strategies [5] are not clinically superior compared to home-based rehabilitation with support. Support system for home-based rehabilitation is required as evident from the patient experiences [22–24] and from the findings of our study. Technology has been used to reduce barriers to access personalized care without increasing treatment burden to the individual. Telerehabilitation or technology-based rehabilitation as a modality was shown to be comparable to face-to face rehabilitation with respect to joint outcomes [25, 26]. However, a cost-effectiveness analysis from Italy revealed standard rehabilitation procedures to be cost-effective compared to Telerehabilitation [27] and another study from Australia [28] concluded that the costs and clinical outcomes were comparable with in-person rehabilitation but significantly reduced carers and patients burden of treatment.

The evidence for use of physiotherapy procedure such as continuous passive motion of operated knee [29, 30], neuromuscular electrical stimulation [31], progressive resistance therapy [32] and aquatic therapy [33] are either inconclusive or lack strong evidence. Hence, the recommendation across countries and health systems may vary but discharge to home and prescription of exercises to perform by self is recommended for non-complicated cases [34] and is also the current recommendation in India [7]. Since lack of self-efficacy and motivation is a major deterrent for adherence, post-surgery monitoring needs to improve in the context of Indian patients. Further research in improving physician–patient communication and better monitoring using technologies should be developed and tested.

In conclusion, large out-of-pocket expenditures are incurred by patients during rehabilitation phase as they find self-rehabilitation challenging. There is a need for easier access to professional support without increasing costs to provide continuing support following discharge during the early months following knee arthroplasty.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Study conceptualisation and design: ND, DG, RM, IVM. Data acquisition: IVM, HD, AK. Data analysis: IVM, ND, SR. Manuscript: First draft IVM and ND. All authors contributed in revising the manuscript and agreed to the final version.

Funding

None.

Declarations

Conflict of interest

The authors declare that they have no conflict of interest.

Ethics Approval

This article does not contain any studies with human or animal subjects performed by the any of the authors.

Informed Consent

For this type of study informed consent is not required.