Impact of vision impairment on discharge destination for patients with hip fracture

Abstract

Introduction

Vision impairment (VI) due to low vision or blindness is a major sensory health problem affecting quality of life and contributing to increased risk of falls and hip fractures (HF). Up to 60% of patients with hip fracture have VI, and VI increases further susceptibility to falls due to mobility challenges after HF. We sought to determine if VI affects discharge destination for patients with HF.

Materials and methods

Cross-sectional analysis of 2015 Inpatient Medicare claims was performed and VI, blindness/low vision), HF and HF surgery were identified using ICD-9, and ICD-10 codes. Patients who sustained a HF with a diagnosis of VI were categorized as HF + VI. The outcome measure was discharge destination of home, skilled nursing facility (SNF), long-term care facility (LTCF) or other.

Results

During the one-year ascertainment of inpatient claims, there were 10,336 total HF patients, 66.82% female, 91.21% non-Hispanic white with mean (standard deviation) age 82.3 (8.2) years. There was an age-related increase in diagnosis of VI with 1.49% (29/1941) of patients aged 65–74, 1.76% (63/3574) of patients aged 75–84, and 2.07% (100/4821) of patients aged 85 and older. The prevalence of VI increased with age, representing 1.5% (29/1941) of adults aged 65–74, 1.8% (63/3574) of adults aged 75–84, and 2.1% (100/4821) of adults aged 85 and older. The age-related increase in VI was not significant (P = 0.235). Patients with HF were most commonly discharged to a SNF (64.46%), followed by ‘Other’ (25.70%), home (7.15%), and LTCF (2.67%). VI was not associated with discharge destination. Male gender, Black race, systemic complications, and late postoperative discharge significantly predicted discharge to LTCF with odds ratios (95%CI) 1.42 (1.07–1.89), 1.90 (1.13–3.18), 2.27 (1.66–3.10), and 1.73 (1.25–2.39) respectively.

Conclusions

The co-morbid presence of VI was not associated with altered discharge destinations to home, skilled nursing facility, LTCF or other setting.

1. Introduction

Millions of older adults fall each year in the US, leading to catastrophic injuries, over 32,000 deaths and healthcare costs of 55 billion dollars annually.^1,2 Many intrinsic and extrinsic factors contribute to fall risk for older adults, including vision impairment (VI). The American Academy of Ophthalmology defines VI as best-corrected visual acuity of less than 20/40 in the better seeing eye. VI includes both low vision, which is VI that cannot be corrected by glasses, surgery or medication, and blindness, which is vision worse than 20/200 or visual field constriction of 20° or less in the better seeing eye.³ VI is a major sensory health problem that impacts quality of life that is an independent risk factor for falls^{4, 5, 6} that significantly affects safe mobility,^{7, 8, 9} results in impaired postural stability and contributes to difficulty navigating environmental hazards and obstacles.¹⁰

With the growing prevalence of VI due to the aging population, VI is a major sensory health problem and has been shown to disproportionately affect minority populations.¹¹ In addition to the impact on falls and mobility, VI is associated with increased risk for fractures secondary to falls, including hip fractures (HF).^12,13 HF can lead to temporary or permanent immobility, acute or chronic pain, limitations in activities of daily living, loss of independence, and mortality for older adults, in addition to associated economic and social burdens for the individual, their family, and the health care system.^{14, 15, 16} The incidence of HF among older adults is expected to increase with the aging population, highlighting this growing public health concerns with VI and falls with the need to improve and ramp up prevention efforts.

Despite awareness that VI increases risk for falls and hip fracture, the impact of VI on post-surgical disposition is not well understood. The purpose of this study was to determine if VI affects discharge destination as an outcome measure for patients admitted for HF. We hypothesized that patients with HF and VI would have greater discharge to skilled nursing facilities compared to patients without VI.

2. Methods

This study was exempt from approval by our Institutional Review Board. A cross-sectional analysis of inpatient Medicare claims from 2015 for individuals aged 65 and older was conducted identifying VI, HF, comorbidities, HF surgery, systemic and surgical complications using ICD-9 and ICD-10 codes. These ICD-9 and ICD-10 codes used were based on previous research^8,17 using data from the Centers for Medicare & Medicaid Services. Vision impairment was defined by ICD-9 diagnosis codes (369.0–369.8, 369.10–369.18, 369.20–369.25, 369.3, and 369.4) and from there ICD-10 codes were derived (H54.0, H54.2, H54.3, H54.8, and H54.10 – H54.12) based on similar diagnoses ¹⁰. Patients who sustained a HF were then categorized as (HF + VI) if they had a diagnosis of VI that included blindness or low vision. Definition of VI is described in a large-scale Medicare population study by Hamedai et al.⁸

The outcome measures were discharge destination of home, skilled nursing facility, long-term care facility (LTCF) or other. The other category included destinations such as psychiatric hospitals, miscellaneous facilities, and internal transfers to receive other services. The Chi-square and Wilcoxon rank sum tests were used to determine univariate association between VI and other variables. Multivariable logistic regression models were developed to study the association between discharge to LTCF and blindness accounting for other covariates. SAS version 9.4 was used to perform all statistical analyses. Statistical significance assessed at alpha of 0.05.

3. Results

During the one-year ascertainment of inpatient Medicare claims, there were a total of 813,341 unique inpatient claims of which 10,336 were HF. This cohort with HF had an average age (standard deviation) of 82.3 (8.2) years, 66.82% female, 91.21% non-Hispanic white (see Table 1 for demographic summary). Of these 1.9% (192) had a diagnosis of VI. The prevalence of VI increased with age, representing 1.5% (29/1941) of adults aged 65–74, 1.8% (63/3574) of adults aged 75–84, and 2.1% (100/4821) of adults aged 85 and older. In this sample size, the age-related increase in VI was not significant (P = 0.235).

Table 1.

Demographics of study population.

	Overall (N = 10336)	Vision Impaired (N = 192)	Not Vision Impaired (N = 10144)	P- Value
Age (years)
65–74	1941 (18.78%)	29 (15.10%)	1912 (18.85%)
75–84	3574 (34.58%)	63 (32.81%)	3511 (34.61%)
≥84	4821 (46.64%)	100 (52.08%)	4721 (46.54%)	0.244
Sex
Female	6907 (66.82%)	128 (66.67%)	6779 (66.83%)	0.963
Male	3429 (33.18%)	64 (33.33%)	3365 (33.17%)
Race
Non-Hispanic White	9427 (91.21%)	177 (92.19%)	9250 (91.19%)	0.812
Black	493 (4.77%)	9 (4.69%)	484 (4.77%)
Other	416 (4.02%)	6 (3.13%)	410 (4.04%)

There was a similar percentage of VI among women with HF at 1.9% and for men with HF at 1.9%. There was no difference between VI by sex (P = 0.963). For the outcome measure of discharge destination, patients who were hospitalized with HF and with HF + VI were mostly discharged to SNF (64.4% HF; 67.2% HF + VI), followed by Other (unspecified) locations (25.8% HF; 22.4% HF + VI) (Table 2). It was rare for either group to be discharged to their home with or without services (7.1% HF; 8.9% HF + VI) or to an LTCF (2.7% HF; 1.6% HF + VI). Overall, there was no statistically significant association between discharge destination and visual impairment, p = 0.43. In the multivariable model male gender, Black race, systemic complications, and late postoperative discharge significantly predicted discharge to LTCF with odds ratios (95%CI) of 1.42 (1.07–1.89); 1.90 (1.13–3.18); 2.26 (1.66–3.09); and 1.73 (1.25–2.39) respectively. VI did not predict discharge to LTCF.

Table 2.

Summary of discharge location.

Discharge Location	Overall (N = 10336)	Vision Impaired (N = 192)	Not Vision Impaired (N = 10144)	P-Value
Home	739 (7.15%)	17 (8.85%)	722 (7.12%)	0.431
LTAC	276 (2.67%)	3 (1.56%)	273 (2.69%)
SNF	6663 (64.46%)	129 (67.19%)	6534 (64.41%)
Other	2658 (25.7%)	43 (22.40%)	2615 (25.78%)

4. Discussion

In this one-year cohort of inpatient Medicare patients admitted for HF, 1.9% of this study cohort had VI and was similar among men and women. All patients with HF, regardless of vision status or sex, were primarily discharged to skilled nursing facilities. Less than 10% of patients were discharged to home for either group, whereas other studies that found 14–20% of hip fracture patients were discharged to home.^18,19 This disparity may be due to differences in coding for discharge destination or due to natural variations observed by year of discharge.¹⁸

Older adults comprise the fastest growing segment of the American population and in conjunction with increases in chronic diseases such as diabetes, and VI is expected to double over the next 30 years.²⁰ The results of this study are similar to previous studies, demonstrating there is a positive correlation between age and VI, with an estimated prevalence of VI, for adults age 40 and older in the United States of 0.9%, ranging from 0.3% for those age 40–49 to 25.7% for individuals who are 80 and older. Other studies looking at the prevalence of VI in patients with HF described the range of VI from 15.4%²¹ to 46%.²² This marked variability may be a consequence of different population sets over time and perhaps undercoding of vision impairment and blindness. In our cohort, the 1.9% VI is similar to U.S. population estimates for blindness for adults age 65 and older (2.8%).²³

Other studies have found that vision impairment can affect discharge location. In a study of older adults (≥75 years old) hospitalized for acute myocardial infarction (AMI), there was an observed difference in discharge location to home for those with VI (81%) versus those with no VI (85%).²⁴ Another study observed that COVID-19 patients with impaired sensory function (vision/hearing/sensation deficits) were more likely to be discharged to an institution (13.7%) rather than home (3.9%).²⁵ Sathiyakumar et al.¹⁹ looked at factors increasing risk for discharge to rehabilitation for patients with HF and included impaired sensorium, and while this likely includes VI, the variable is not defined in the study. However, they did not find a significant association between impaired sensorium and discharge location. The current study found 67% of vision impaired hip fracture patients were discharged to a SNF, which was similar to Sathiyakumar et al.’s finding of 78% of hip fracture patients with impaired sensorium discharged to a rehabilitation facility.

Patients with HF experience fine motor skill and mobility functional declines post-operatively demonstrated by declines in activities of daily living, instrumental activities of daily living, large motor, gross motor, and cognitive status scores.¹⁸ This multifactorial context makes recovery without assistance difficult and may contribute to risk for falls or further injury. Impaired vision alone is associated with decreased mobility⁹ and independence as well as vision-specific functioning (e.g., difficulty seeing stairs even with corrective eyewear),^26,27 in addition to increased fall risk,¹⁰ which could make individuals with HF + VI vulnerable to additional injury. Acute care settings are recommended for older adults with VI during occupational therapy to maximize rehabilitation efficacy and enhanced screening of older adults who are a higher risk for difficulty with discharge instructions and appropriate use of medications.^28,29 Yet the findings of this study suggest that there was no difference in discharge destination for hip fracture patients with and without vision impairment. It is possible that hip fracture patients, who in general are more vulnerable, are more likely to be referred to an acute care setting than individuals hospitalized for other reasons.

In our sample, we were surprised that VI was not associated with different discharge outcomes for patients with HF, but this may be due to the limited sample size based on a single year. A possible explanation is that many individuals with VI may have an existing support system that allows them to return home at the same rate as individuals without VI. Such a possible explanation is feasible given a prior study that demonstrated that an individual's social support network size is a predictor for whether visually impaired older adults will undergo certain medical procedures.³⁰

Other potential limitations of this study include under coding for blindness or low vision in sampling. Additionally, this study focused on a cross-sectional cohort without analysis of change over time. Nearly 2% of the study population was coded as visually impaired, which is similar to general population estimates for Americans aged 65 and older.³¹ However, both ICD-9 and ICD-10 codes were used due to the Medicare transition during the study period. This ICD transition impacts the number of people included in our cohort due to a change in how ICD codes identified VI.

5. Conclusion

This study looked at the outcome of discharge destination for patients with HF aged 65 and older and the impact of VI on this outcome. There was no difference between discharge destination for patients with HF with or without VI or by sex. Future research could examine the post-operative outcomes of patients with HF and effect of VI in relation to additional measures of the social determinants of health (SDOH) or varying degrees or specific pathologies of visual impairment. The addition of such data on SDOH will assist in data patterns and modeling to improve transitions of coordinated care in both immediate and long-term care for patients that sustain hip fractures.

Declaration of conflicting interests

The authors declare that there is no conflict of interest.

Funding

This work was supported by the National Institute on Aging under the GEMSSTAR (Grants for Early Medical and Surgical Subspecialists’ Transition to Aging Research) grant (R03AG060177-02) and the Paul B. Beeson Emerging Leaders Career Development Award in Aging (K76AG068435).

Ethical statement

This study was exempt from approval by our Institutional Review Board.

CRediT authorship contribution statement

Jacarri Tollette: Conceptualization, Methodology, Formal analysis, Investigation, Data curation, Writing – original draft, Writing – review & editing. Victor Heh: Conceptualization, Methodology, Formal analysis, Writing – review & editing. Jessica M. Wiseman: Conceptualization, Methodology, Formal analysis, Investigation, Data curation, Writing – review & editing, Supervision, Project administration. Catherine C. Quatman-Yates: Conceptualization, Methodology, Validation, Formal analysis, Writing – review & editing. Sayoko Moroi: Conceptualization, Methodology, Validation, Formal analysis, Writing – review & editing. Carmen E. Quatman: Conceptualization, Methodology, Validation, Formal analysis, Writing – review & editing, Supervision, Funding acquisition.