Skip to main content

Table 1 Studies assessing the impact of obesity on resource use in total hip or knee arthroplasty

From: Addressing obesity in the management of knee and hip osteoarthritis – weighing in from an economic perspective

Author, Year Healthcare setting/Country Study population Research aim/focus Measurement of obesity Costing perspective/measurement and types of counted Results
Epstein AM, et al, 1987 [25] Large acute care hospital, USA 278 patients who underwent TKA and 111 patients who underwent THA, October 1983 -September 1984. To determine the relationship of body weight to LOS and total charges for all patients undergoing THA or TKA Height and weight taken from pre-operative medical records.
5 levels of weight status categorised by actual weight compared to ideal weight as a %
Health service provider perspective capturing charge data for the inpatient stay only. Extremely overweight patients (≥ 188 % ideal) had 35 % mean longer LOS (p < 0.01) and 30 % higher total charges (p < 0.01) than normal weight counterparts. Extremely underweight patients also reported significantly higher costs.
Jibodh SR, et al, 2004 [26] Large acute care hospital, USA 188 patients who underwent primary THA, 1996 – 2001. To determine the influence of BMI on perioperative morbidity (time of surgery until discharge) on functional recovery and hospital service use (LOS,total and individual cost items) Height and weight taken from pre-operative medical records to calculate BMI and categorised into non obese(BMI < 25), mild(BMI >25-29.9), moderate (BMI>30-39.9) and severe (BMI >40) Health service provider perspective using hospital charge data and reporting total charges and 8 separate billing categories No significant difference in LOS between 4 BMI groups. A trend towards higher overall charges with increasing obesity but not statistically significant. No significant differences in any of the individual charges were noted between 4 BMI groups in any of 8 billing categories, however morbidly obese patients longer mean operative time (P < 0.05)
Vincent HK,et al, 2007 [27] Inpatient rehabilitation hospital, USA 342 participants who underwent primary or revision TKA, January 2002 - March 2005. Complete case analysis on 285 participants. To examine the effect of obesity on functional and financial outcomes in patients with TKA undergoing inpatient rehabilitation. Height and weight taken from patient medical records to determine BMI and categorised as obese (BMI > 30) or non-obese(BMI < 30) Health service provider perspective using hospital charge data collecting total hospital charges and daily charges for period of inpatient stay only. LOS was longer in primary and revision obese patients (9.8 days) than for non- obese patients (8.8 days) (P < 0.05). Total charges were higher for obese patients (USD 12,386) than non -obese patients (USD 10,618) (P < 0.005). Primary TKA group; total hospital charges were significantly higher in the obese than non- obese group (P < 0.05)
Vincent HK et al, 2007 [28] Inpatient rehabilitation hospital, USA 339 obese and non- obese patients with primary or revision THA, January 2002- March 2005. Complete case analysis on 178 participants. To examine the effect of increasing BMI on functional and financial outcomes in patients with THA undergoing inpatient rehabilitation Height and weight taken from patient medical records to determine BMI and categorised as non- obese (BMI < 25) overweight (BMI 25-30) obese (BMI > 30-39.9) and severely obese (BMI ≥ 40) Health service provider perspective collecting total hospital charges and daily charges (using total charges and dividing by LOS) for the period of the inpatient stay only. LOS were significantly different in the severely obese group compared with the non- obese group (p < 0.05). A significant curvilinear relationship between LOS and BMI with the lowest LOS found in overweight and obese persons (R squared =0.124 P < 0.05). Total charges were greater in the severely obese group compared to the overweight group (P < 0.05).
Vincent HK & Vincent KR, 2008 [29] 15 independent rehabilitation hospitals, USA 5428 obese and non-obese patients who underwent primary TKA or revision TKA, January 2002- March 2006. To determine the influence of obesity on rehabilitation outcomes including LOS and hospital charges following TKA Height and weight taken from patient medical records to determine BMI and categorised as non-obese (BMI < 25) overweight (BMI 25-30) obese (BMI > 30-39.9) and severely obese (BMI ≥ 40) Health service provider perspective with collection of total charges and pharmacy, occupational and physical therapy rehabilitation hospital charges LOS was longest in the non- obese group compared to all other groups (P < 0.05) but age differences amongst groups likely to be impacting on results. The severely obese group had the highest daily charges (USD 36 excess dollars) (p < 0.05) but not physical therapy charges or total charges which was highest in the non -obese group (P < 0.05). A significant interaction effect was found for TKA status (primary versus revision) and BMI group for total charges (P < 0.05).
Batsis JA, et al, 2010 [30] Large acute care hospital, USA 5539 uncomplicated TKA recipients, 1996- 2004 and classified by BMI (WHO) categories. To determine the impact of BMI on post-operative outcomes and resource utilization following elective TKA Height and weight taken at time of surgery and recorded in own joint registry to determine BMI and categorised as BMI normal (BMI 18.5- 24.9) overweight (BMI 25-29.9) obese (BMI > 30-34.9) and morbidly obese (BMI ≥ 35.0) Health service provider perspective with all direct costs associated with inpatient stay including physician services and readmission within 30 days associated with the primary surgery. Overall costs were similar among normal, overweight, obese or morbidly obese patients (P = 0.24) Post-surgical costs were no different among groups (P = 0.44). Higher BMI was associated with a higher mean anaesthesia and operative times and a higher overall Charlson comorbidity index.
Batsis JA, et al, 2009 [31] large acute care hospital, USA 5642 unilateral uncomplicated THA patients between 1996 -2004 and classified by BMI categories. To determine the impact of BMI on post-operative outcomes and resource utilization following elective THA Height and weight taken at time of surgery and recorded in own joint registry to determine BMI and categorised as BMI normal (BMI 18.5- 24.9) overweight (BMI 25-29.9) obese (BMI > 30-34.9) and morbidly obese (BMI ≥ 35.0) Health service provider perspective with all direct costs associated with inpatient stay including physician services and readmission within 30 days associated with the primary surgery. No significant differences between BMI groups for LOS, post-operative overall, hospital and physician costs. Operative and anaesthesia costs were higher in morbidly obese group than all other groups. All other adjusted costs were non-significant. No significant differences between groups in: composite 30 day endpoints, rate of patient transfers to ICU or number of days in ICU.
Kim, SH, 2010 [32] Short stay, community hospitals in the Nationwide Inpatient Sample (NIS- 2006), USA 229 001 primary TKA recipients and 497 001 primary THA recipients in the USA captured in the NIS. To estimate the prevalence of morbid obesity (≥40 kg/m2 in the THA and TKA sample and to determine if there is greater resource use attributable to morbid obesity for primary TJA Presence of obesity (BMI ≥30.0) and morbid obesity (BMI ≥ 40.0) identified by the corresponding ICD_9M codes for obesity in hospital administrative databases Health service provider perspective using hospital inpatient charge data converted to cost data and reporting on overall hospital costs only When adjusted for known confounders, hospital resource consumption for primary THA and TKA was 9 % and 7 % higher among morbidly obese than among non-obese patients respectively
Dowsey M, et al, 2011 [33] Large acute hospital, Australia 521 primary TKA recipients, January 2006 - December 2007. To determine whether obesity was independently associated with higher hospital costs for the index procedure and over the following 12 months. Presence of obesity (BMI ≥30.0) captured from preoperative measures recorded in own hospital joint registry Healthcare service provider perspective capturing total inpatient costs for the index TKA, relevant readmissions in the first 12 months and the two together named episode of care. Statistically significant association between obesity and higher inpatient costs ($1127 P = 0.036) and higher episode of care costs (+1,821 P = 0.024). Using BMI as a continuous variable, cost of index procedure increased by $129 and episode of care costs increased by $159 per unit increase of BMI.
Silber JH, et al, 2012 [34] 47 acute hospitals of varying size across multiple locations, USA 2045 obese patients (BMI ≥ 35 kg/m2) matched to non-obese patients undergoing THA, TKA (primary or revision), colectomy, thoracotomy, 2002- 2006. 75 % of the sample underwent TJA. To study the medical and financial outcomes associated with surgery in the elderly obese. Presence of severe obesity (BMI ≥ 35.0 < 40.0) and morbid obesity (BMI ≥ 40.0) captured from baseline BMI data in hospital medical records Healthcare service provider perspective using 2 alternate costing methods (Medicare payments versus costs using cost to charges ratios) (to determine overall hospital costs from admission to 30 days post operation. Medicare payments were 3 % greater (P < 0.001) and provider costs were 10 % greater for obese compared to non- obese matched counterparts (P < 0.001). The Obese group recorded a 12 % longer LOS than their complete matched non obese counterparts (P < 0.001)
Maradit Kremers H, et al, 2014 [35] Large acute care hospital, USA 8129 patients who underwent 6475 primary TKA and 1654 revision TKA, January 2000 - September 2008. To examine the relationship between obesity, length of stay and direct medical costs during the index hospitalisation and a 90 day window taking into account obesity related co-morbidities. Height and weight taken from patient admission records to calculate BMI and categorised into 8 BMI categories and as a continuous variable Health service provider perspective using hospital administration databases and converting charges to costs using cost centre specific ratios. End points of hospital LOS, direct medical costs during hospitalisation and total medical costs during the 90 day window LOS was longer at the extreme ends of the BMI spectrum only with mean LOS lowest in those with BMI 30-40.0.
After adjusting for known confounders, every 5 unit increase in BMI over 30 was associated with higher mean costs of USD 421 for hospitalisation and USD 524 for 90 days and remained significant after adjustment for comorbidities (P = <0.001) and complications (P = 0.004).
Maradit Kremers H, et al, 2013 [36] Large acute care hospital, USA 8973 patients; 6410 primary THA and 2563 revision THA's, January 2000 - Sept 2008. To examine the relationship between obesity, length of stay and direct medical costs during the index hospitalisation and a 90 day window taking into account obesity related co-morbidities. Height and weight taken from patient admission records to calculate BMI and categorised into 8 BMI categories and as a continuous variable Health service provider perspective using hospital administration databases and converting charges to costs using cost centre specific ratios. End points of hospital LOS, direct medical costs during hospitalisation and total medical costs during the 90 day window Increasing BMI was associated with higher hospital costs and this association persisted among patients without significant comorbidities or complications. After adjusting for known confounders, every 5 unit increase in BMI was associated with USD 744 and USD 1183 higher hospitalisation and 90 day costs respectively. (This corresponds to about 5 % higher hospitalisation and 90 day costs respectively).
  1. TKA total knee arthroplasty, THA total hip arthroplasty, TJA+ total joint arthroplasty, LOS length of stay, BMI body mass index, USD USA dollars, NIS national inpatient survey, ICD-9 M The International Classification of Diseases, 9th Revision