Sex and Age Effects on Outcomes of Total Hip Arthroplasty After Inpatient Rehabilitation

Study Design 

The design was an exploratory, retrospective study utilizing data from the computerized medical records system of the university-affiliated rehabilitation hospital. We collected and analyzed the data for this investigation in accordance with the procedures and policies set forth by the Human Investigation Committee at the University of Virginia (UVA).

Participants 

The population sample was a convenience sample. Data from the UVA Clinical Data Repository indicated that the total UVA THA population during the time frame of 2002 to 2005 consisted of 697 cases, of which 42% were men and 58% women. Fifty percent of the patients were less than 65 years, 34.5% were between 65 and 80 years, and 14.5 were more than 80 years. Most of the patients were white (84.94%) and African American (13.92%). The average acute-care LOS before rehabilitation was 4.38 days, with nearly identical LOS for men and women (4.19d and 4.52d, respectively), with similar LOS for all age brackets. At discharge, 12% of patients were routinely discharged home, 43% were discharged to inpatient rehabilitation facilities, and the remaining patients were discharged directly to skilled nursing facilities (SNFs), home care, or home health services.

We identified 332 patients with diagnoses of THA from the computerized database between the dates of January 1, 2002, and March 31, 2005. Among these patients, 193 had primary THA and 139 had revision THA. We stratified patients into 1 of 3 age groups based on the reimbursement divisions established by the Center for Medicare Services (<65y, 65–84y, ≥85y). All completed the THA procedure with 1 of 3 orthopedic surgeons specializing in joint arthroplasty at this institution. All participants were typically admitted to the inpatient rehabilitation hospital 3 to 5 days after the surgery. Patients were provided a standard pain control regimen before and during the rehabilitation stay by the health care team to reduce the potential effects of pain on rehabilitation progress.

Population Variables 

Patient descriptive variables included age, sex, and the number and type of comorbidities. The total number of comorbidities was consolidated for each patient at the time of admission and presented as a total number. Other descriptors included height, weight, body mass index (BMI), smoking status, current alcohol use, and living alone. The use of a hip abductor brace was noted for these patients because this can affect functional outcomes. Finally, we recorded the THA surgery type (primary or revision) and revision diagnoses (pain, mechanical, infection).

Inpatient Rehabilitation Program 

All patients completed comprehensive, interdisciplinary inpatient rehabilitation programs at the UVA Rehabilitation Hospital (up to 2wk). Each patient received approximately 3 hours of supervised therapy daily from both physical and occupational therapists. The general rehabilitation program was conducted under the supervision of a physiatrist. Therapy sessions were conducted 2 times daily (morning and afternoon sessions) and consisted of both individual and group sessions. During therapy sessions, patients performed activities designed to improve flexibility and range of motion, improve independence with performing ADLs, and to improve gait, balance, and proprioception. Patients were further instructed on the use of assistive devices for ambulation as dictated by the patient’s individual needs, and all patients were placed on hip precautions on the operated side limiting hip flexion to less than 90°, no internal rotation, and no adduction past midline. Further therapeutic interventions from speech pathology or pain psychology were provided on a case-by-case basis if necessary.

Outcome Variables 

Functional improvement during inpatient rehabilitation is commonly measured using the 18-item FIM.22 This instrument is composed of cognitive and physical domains, which are evaluated in each patient by the rehabilitation team. The FIM estimates performance during tasks that can be broadly categorized as activities of daily living, mobility, and cognitive domains, and it is an important representative measure of the success of an interdisciplinary rehabilitation program. Given that the primary goal identified by almost all THA patients on their program plan was to “get home” and do “normal activities,” the FIM was a validated measure to capture overall functional improvements. The total possible score on the FIM is 126, and THA patients have previously shown more than 20-point improvements.9, 12, 22

The LOS, FIM scores at admission and discharge, total rehabilitation hospital charges, and discharge disposition were the variables collected from the computerized database and individual patient charts. We adjusted total charges for the LOS to generate a cost-per-day variable. As an estimate of the functional gain during the rehabilitation intervention, a FIM efficiency score was reported. FIM efficiency was calculated as the change in FIM from admission to discharge divided by the LOS (FIM points gained/LOS in days).

Statistical Analyses 

We performed statistical analyses using the SPSS software.a All data are expressed as mean ± standard deviation (SD). A 1-way analysis of variance (ANOVA) was used to determine whether differences existed between men and women for baseline characteristics and descriptive characteristics. We used chi-square tests to determine whether frequencies of characteristics (eg, smoking status, current alcohol use, living alone, and hip brace use), discharge disposition locations, reasons for revision surgeries, and weight bearing and joint replacement status were different among age brackets or sex.

The interaction between age and sex on the major outcome variables was analyzed using a univariate ANOVA with Tukey post hoc tests to determine where differences existed. A repeated-measures ANOVA was performed on FIM from admission to discharge. The dependent variable was FIM, with the group factor (sex, age, THA type) and time factor (admission, discharge). A Tukey post hoc test was used to determine where differences occurred. Because there was a significant difference among groups for admission FIM scores, we performed an analysis of covariance (ANCOVA) to evaluate discharge FIM scores for the age and sex comparisons. We used admission FIM scores as the covariate and the predicted means generated by the ANCOVA were analyzed using a 1-way ANOVA with a Tukey post hoc test. Given that our previous work has shown that THA type (primary vs revision) influences THA outcomes, we examined the potential interactions of age, sex, and THA status of the 2 main outcome variables FIM efficiency and LOS by applying a univariate ANOVA with group factors of age, sex, and THA status in this exploratory study. Because of unequal sample sizes in these group stratifications, results were corroborated using nonparametric tests (Mann-Whitney U tests). Given that the findings were the same for each form of analysis, the ANOVA analyses are presented in the results section. A priori α levels were set at .05 for all statistical tests.

Participant Characteristics 

Table 1 contains the subject characteristics for men and women. More men were heavier, were smokers, and had ischemic heart disease than women (P<.05). More men (77.75%) were insured by Medicare, whereas 17.6% were insured by private carriers. Most of the women (84.1%) were insured by Medicare and 8.5% were insured by private carriers (not shown). All other descriptors and revision etiologies were not statistically different between men and women. The comorbidities are shown in table 2. A greater percentage of women lived alone and had hypertension and hypothyroidism than men (P<.05). More women were weight-bearing-as-tolerated status (reflecting cemented joint components) and fewer were partial weight-bearing status (reflecting noncemented joint components) than men (P<.05).

Table 1. Subject Characteristics of Male and Female THA Patients (N=332)

	Characteristics	Men (n=121)	Women (n=211)
	Age (y)	69.2±10.6	71.4±10.2
	Height (cm)	144.1±23.5	140.3±22.2
	Weight (kg)	87.2±20.3	75.9±22.2⁎
	BMI (kg/m2)	30.1±6.2	29.5±7.9
	BMI over 30kg/m2 (% of group)	39.4	44.6
	Smoker (% of group)	16.3	10.6⁎
	Living alone (% of group)	11.4	16.1⁎
	Race (% of group)
	African American	11.3	14.5
	White	88.8	84.7
	Other	0	0.8
	Use of hip abductor brace (% of group)	7.5	6.1
	Primary, revision (% of group)	61.7, 38.3	56.5, 43.5
	Revision diagnosis (% of group)
	Mechanical	20.8	22.0
	Infection	3.3	6.2
	Pain	14.2	15.3
	Weight-bearing status (% of group)
	Weight bearing as tolerated	41.8	63.1⁎
	Partial weight bearing	57.0	36.9⁎
	Non–weight bearing	1.3	0.0

NOTE. Values are mean ± SD or as indicated.

⁎ Different from men at P<.05.

Table 2. Comorbidities of Male and Female THA Patients

	No. of Comorbidities	Men	Women
		7.3±1.5	7.3±1.6
	Type of comorbidities (% of group)
	Diabetes mellitus	17.0	18.0
	Ischemic heart disease	20.7	15.6⁎
	Heart failure	6.6	7.1
	Hyperlipidemia	24.0	22.7
	Hypertension	62.8	71.6⁎
	Hypothyroid	8.3	19.0⁎
	Anemia	40.5	40.3
	Rheumatoid arthritis	5.0	6.6
	Gout	3.3	0.5
	Depression	10.7	12.8
	General debility	40.5	37.0

NOTE. Values are mean ± SD or as indicated.

⁎ Different from men at P<.05.

FIM Scores 

FIM scores are presented by age and sex in table 3. Admission FIM scores were 14% lower for the women than the men (P=.003). Admission FIM was lowest for the patients older than 85 years (P=.004). Men had 15% higher discharge FIM scores than women (P=.003). Patients 65 to 84 years had higher FIM scores than the 2 other age groups (P=.007). Discharge FIM scores were 12% higher in patients with primary THA than those with revision THA (P<.001). The interaction term of age, sex, and THA type was not statistically significant (F2=2.473, P=.086). All patients increased FIM scores from admission to discharge regardless of age and sex. Neither age nor sex was a significant main effect for FIM change scores (both P>.05). The age and sex interaction term was not significant (P=.260).

Table 3. FIM Instrument Scores at Admission and Discharge After Inpatient Rehabilitation for Primary and Revision THA Patients

	Subjects	Admission (points)	Discharge (points)	FIM Change (discharge – admission)
	Men
	<65y (n=35)	82.1±9.3	107.1±9.2	24.9±7.4
	65–84y (n=75)	80.1±8.4	109.3±14.2	29.2±14.5
	≥85y (n=10)	73.6±7.0	103.6±1.5	30.0±7.1
	Women
	<65y (n=41)	76.1±8.6	102.9±8.2	26.8±9.1
	65–84y (n=159)	75.8±10.5	104.9±12.8	29.1±9.1
	≥85y (n=11)	67.8±12.8	91.0±12.5	23.2±4.6
	Univariate ANOVA F Test
	Sex (P)	.003⁎	.003⁎	.355
	Age (P)	.004⁎	.007⁎	.067
	Age by sex (P)	.799	.416	.260

NOTE. Values are mean ± SD.

⁎ Significant effect at P<.05.

Length of Stay 

Age and sex analyses for LOS are shown in figure 1. LOS was longer in women than men regardless of age (11.7d vs 9.6d; P=.012). Patients in the 3 age brackets did not have statistically different LOS (F=1.640, P=.196). The THA type effect did not reach statistical significance (P=.065), but the power was low (P=.45), and the interaction term was not significant (P=.43; observed power, .192) due to a small cell size with the LOS group fractionations.

View Large Image Download to PowerPoint

Fig 1. LOS in THA patients presented by age and sex. Values are means ± SD. *Greater than men at P<.05.

FIM Efficiency 

The overall FIM efficiency was 26% lower in women than men regardless of age (P=.006; fig 2A). The main effect for age was significant, with patients greater than 85 years having lower FIM efficiency scores than patients 65 to 84 years (P=.015). FIM efficiency scores were 35% higher in the primary THA than revision THA (P=.016). FIM efficiencies are shown in figure 2B.

View Large Image Download to PowerPoint

Fig 2. (A) FIM efficiency scores presented by age and sex. Values are mean ± SD. *Significantly different from men in same age group (P<.05); †significantly different from women in other age groups (P<.05). (B) FIM efficiency scores presented by age and sex and THA type. Values are mean ± SD. *Significantly different from men in same age and THA type group (P<.05); †significantly different from primary THA in same age and sex group (P<.05).

Total Rehabilitation Hospital Charges 

Overall, total rehabilitation hospital charges were 17% higher for women than men (P=.003; table 4). Similarly, daily rehabilitation hospital charges were significantly greater (4%) for women than men. The total and daily hospital charges for patients 85 years or older were higher than those for patients 65 to 84 years (P<.05). For women and men, patients 85 years or older had higher total charges than those less than 65 and 65 to 84 years (P<.05).

Table 4. Inpatient Rehabilitation Hospital Charges for Male and Female THA Patients in 3 Age Brackets

	Patients	Total Charges ($)	Charge per Day ($)
	Men
	<65y	11,331±1,864	1,172±442
	65–84y	10,771±4,614	1,126±138
	≥85y	14,346±2,834	1,127±148
	Grand mean, men	11,141±1,136	1,140±263
	Women
	<65y	12,662±6819	1227±275
	65–84y	13,123±5757	1158±194
	≥85y	14,697±3315	1344±105
	Grand mean, women	13,099±5920⁎	1181±212⁎
	Grand mean, age bracket
	<65y	12,058±1371	1202±366
	65–84y	12,344±5560	1147±177
	≥85y	14,553±4047†	1255±120†

NOTE. Values are mean ± SD.

⁎ Different from total men at P<.05. † Different from 65- to 84-year age bracket at P<.05.

Discharge Disposition 

The percentage of men discharged to home was greater than for women (P<.001; table 5). The proportions of persons discharged to an SNF, back to the acute care hospital, and to assisted living did not differ. When analyzed by age bracket, the lowest proportion of patients discharged to home was the 85 years or older group. A greater proportion of patients 85 years or older were discharged to an SNF and assisted living than patients less than 85 years (P=.05).

Table 5. Discharge Disposition Based on Sex and Age Bracket in THA Patients

	Disposition	Men		Women
	Home	90.9		84.4⁎
	SNF	4.1		8.1
	Acute transfer back to hospital	4.1		7.1
	Assisted living	0.8		0.5
		<65 Years	65–84 Years	≥85 Years
	Home	90.7	86.9	75.0†
	SNF	4.0	6.8	15.0†
	Acute transfer back to hospital	5.3	5.9	5.0
	Assisted living	0	0.4	5.0†

NOTE. Values are percentages of the total population.

⁎ Different from men at P<.05. † Different from patients less than 65 years and 65 to 84 years at P<.05.

Effect of Age on Functional Outcomes 

The elderly patients (≥85y) had lower FIM scores and efficiency than younger counterparts, and this pattern was consistent for men and women. However, elderly persons made similar percentage changes in functional improvement, but at a slower rate than younger counterparts. These data suggest that older and younger adults derive benefit from inpatient rehabilitation, and age should not preclude participation in an inpatient rehabilitation program.

These findings are similar to those found in the Uniform Data System for Medical Rehabilitation National Database (UDSmr, a subscription database available for rehabilitation professionals). Average FIM increased 28 points in this national cohort, with the average FIM of 3.37. The FIM efficiency scores in the patients under 65 years from this study are comparable to those reported by the UDSmr, whereas the elderly THA patients have considerably lower FIM scores. One age-related possibility for lower functionality is state trait anxiety and perceived pain following surgery. Older adults might perceive pain as a signal of unhealed damage to the joint23; pain itself (regardless of the strength of the pain signal) might caution older adults against exerting themselves as much on specific rehabilitation activities, thereby slowing the rate of progress. Although deficits in presurgical preparation and postsurgical expectations (poor coping skills or low life satisfaction) may have existed among the older age groups to contribute that contribute to rehabilitation recovery,24 we were unable to obtain this type of qualitative data. Identification of mental perceptions and barriers during rehabilitation is an important topic for future research and intervention.

LOS was lengthier among older adults 85 years and older than those less than 85 years. Older THA patients were functionally worse than younger patients at admission. More assistance and recovery time was required to get these older adults to obtain a meaningful functional improvement than younger adults. Presurgical characteristics of the older group such as other bodily joint limitations, tolerance to pain, or deconditioning might have contributed to slower recovery during rehabilitation. Lin and Kaplan11 reported that LOS increased linearly with advancing age, and that race (being African American) contributed to lengthy LOS. In this cohort, aging itself may be associated with more complex comorbidities, such as diabetes. There was a progressively higher prevalence of diabetes, ischemic heart disease, and anemia with each age bracket in this study, but there were no differences in racial distribution among brackets.

Influence of Sex on Functional Outcomes 

Data are scarce regarding sex effects on inpatient rehabilitation after THA, but studies report that being female is a risk factor for poor mobility 9 months to 15 years after hip arthroplasty.3, 25 Women are more likely to ambulate a shorter distance and require more assistance with various ADLs than men 1 year or more after THA.14 In this study, women began and ended the short-term rehabilitation course with lower FIM scores than men, and this could be due to lower limb strength and muscle tone.26 Related factors may include a greater prevalence of sedentary behavior in women than men14 that could slow rehabilitation gains. Also, hip dysplasia (a measure we did not collect) occurs more frequently in women and might have compromised weight-bearing tolerance and ROM activities (and FIM)27 by placing excessive stress on the healing tissues.

Sex as the sole factor may not fully explain outcome differences in the present study. Women were an average of 2.2 years older than men and may have had longer durations of damaging coexistent comorbid complications (ie, endocrine disorders, inflammatory disease).3, 28 Women had more diagnoses of hypothyroidism, obesity, and hypertension than men in this cohort. If hypothyroidism was not adequately controlled, fatigue may have ensued to slow rehabilitation progress. Obesity-related pressure on the surgical area may be initially more painful for the patient during rehabilitation, and may be another contributor to longer LOS and FIM efficiency3 in women. Wang et al29 showed that hypertension itself increased LOS and compromised daily functional gains in arthroplasty patients, although the mechanism to explain this was not clear. Further investigation is required to determine what type, duration, and number of sex-related comorbidities influence inpatient rehabilitation outcomes.

Revision THA Effect 

THA, particularly revision THA, is complex, traumatic and causes significant blood loss.30 Inpatient rehabilitation is also complex after THA. The average LOS for our THA patients was 10.7 days, similar to that reported in the literature and UDSmr.12, 31, 32 More women (5.2%) underwent revision surgeries than men, and these women were older. Women likely underwent THA at a more advanced stage of arthritis or underwent revision procedures than men thereby requiring a longer LOS.13 Elderly patients likely deferred THA until pain and immobility could not be tolerated, as has been reported previously.33

Revision status influenced the FIM score regardless of age or sex for many of the outcome variables. Revision THA patients might be more cautious, anxious, and tentative during postoperative therapy because of the fear of another potential joint replacement failure, leading to lower FIM gains during rehabilitation.34 Elderly women with revision fared the worst in this study; FIM efficiency was approximately 9% to 20% of that achieved by the remaining groups. Tissue disruption with a revision surgery (removal of the old components and replacement with new joint prosthetic components) and complication rates may be greater than in primary surgical procedures,15 which could lead to longer LOS and less functional gain. In this study, the use of a hip abductor brace to stabilize the hip joint did not improve or worsen the inpatient rehabilitation performance compared with no brace in either primary or revision groups.

Total Rehabilitation Charges 

Total and daily hospital charges were higher for women than men, and higher for patients 85 years and older than those 65 to 84 years. Similar to recent work, comorbidities associated with specific age brackets or sex, such as anemia, obesity, rheumatoid arthritis, and diabetes, increased the daily and total charges. Charges in these groups were higher most likely because of postsurgical complications (eg, blood loss and anemia),35 medication-related costs (eg, diabetes or hypertension medications), and increased LOS.36

Discharge Disposition 

The goals of inpatient rehabilitation are to restore physical function and independence and to discharge patients to the least restrictive environment. Women and the elderly were discharged least frequently among all groups. These findings are in agreement with Epps,37 who reported that both older and female THA patients were discharged more often to another facility after surgery than their younger and male counterparts. Jones et al38 reported a lower rate of home discharge in THA patients 80 years and older than those less than 80 years (29% vs 59%), a greater number of whom were living alone. In this study, more women and persons 85 years old lived alone than men. This may contribute in part to the inpatient rehabilitation team’s hesitancy to discharge the patient home with no support. Given that more women had lower function and were more likely to live alone, the inpatient rehabilitation team might have determined that the best course of action was to discharge these older women to an SNF or assisted living where ADL skills could be improved within a safe, staffed environment before returning home. Alternatively, complications such as infection in some patients required acute transfer back to the hospital independent of age or sex.

Study Limitations and Strengths 

From this exploratory investigation, we found several patterns of functional improvement among age and sex brackets in this convenience sample. With a retrospective study, mechanisms explaining differences among groups could not be determined, only speculated. Information regarding socioeconomic status, time course of development of joint pain, and osteoarthritis before elective THA, acute care complications, joint component characteristics, specific inflammatory laboratory values, individual patient psychologic outlook before admission, and pain coping skills, may explain differences in rehabilitation outcomes. This study did not attempt to match all subjects and control for all possible confounders such as presurgical functional status and personality traits that can influence recovery processes.

Future research directions for retrospective studies warrant a complete patient review of disease history, surgical procedures, and laboratory workup in addition to inpatient rehabilitation outcomes. Prospective studies may include comparisons of age and sex brackets with interventions to improve FIM efficiency during inpatient rehabilitation after THA. In addition, strategies that optimize muscle strength and physical function (prehabilitation techniques) before the THA should be investigated by age and sex to enhance FIM efficiency and lower hospital charges.

Strengths of this study include that the population sample contained all eligible THA patients with no selection or bias toward a specific subset of THA patients. By including all possible patients (with the variation that comes with this population), the large sample size reflective of the national THA population provided the first look at sex and age effects on several rehabilitation outcomes that can be generalized to other THA populations.