2.1 Data source

This cross-sectional study utilized data from the 2016 to 2021 National Inpatient Sample (NIS), part of the Healthcare Cost and Utilization Project (HCUP). The NIS is the largest publicly available all-payer inpatient database in the United States and is designed to produce nationally representative estimates.

2.2 Participants and setting

We identified hospitalizations involving adult patients who underwent cervical or lumbar spinal fusion procedures using International Classification of Diseases, 10th Revision, Procedure Coding System (ICD-10-PCS) codes (Appendix A). Then hospitalizations were included if the indication for surgery was either trauma or degenerative spinal conditions. Cases were excluded if spinal fusion was not performed or if patients had a diagnosis of malignant neoplasm. After applying inclusion and exclusion criteria, the final sample included 338,578 hospitalizations (weighted 1,692,889 hospitalization).

The primary exposure of interest was legal blindness, identified using relevant ICD-10-CM diagnosis code H548. Patients were classified as either legally blind group or non-blind group (control group). The study included three binary outcomes: (1) major medical complications (MMC), defined as a composite of pneumonia, myocardial infarction, sepsis, acute respiratory distress syndrome (ARDS), stroke, pulmonary embolism, deep vein thrombosis, acute kidney injury, and aspiration pneumonitis; (2) non-home discharge (NHD), defined as discharge to a skilled nursing facility, intermediate care facility, short-term hospital, another type of facility, against medical advice, or death during hospitalization; and (3) extended length of stay (ELOS), defined as greater than three days.

A range of covariates were included in the analysis based on clinical relevance and statistical associations observed in bivariate testing. These included age (continuous), sex (male or female), race (White, Black, Hispanic, Asian/Pacific Islanders, Native Americans or Others), median household income quartile (0–25th, 26–50th, 51–75th, 76–100th), primary payer (Private insurance, Medicare, Medicaid, Self-pay or other), obesity (yes/no), hypertension (categorized as none, uncomplicated, or complicated), congestive heart failure (yes/no), hyperlipidemia (yes/no), trauma status (yes/no), fall history (yes/no), fragility fracture history (yes/no), and hospital teaching status (urban teaching, urban non-teaching, rural).

2.3 Statistical method

All analyses accounted for the complex survey design of the NIS by incorporating discharge weights, strata, and primary sampling units using survey-weighted procedures. Chi-square tests (for categorical variables) and t-tests (for continuous variables) were used to assess differences between the legally blind group and the control group (Table 1). Multicollinearity was assessed using variance inflation factors (VIFs), and no significant collinearity was found among covariates (mean VIF = 1.12). Weighted multivariable logistic regression models were used to assess the association between legal blindness and each outcome of interest, adjusting for the full set of covariates (Tables 2–4). Adjusted odds ratios (aORs) with 95 % confidence intervals (CIs) were reported. The predictive performance of each model was evaluated using the area under the receiver operating characteristic (ROC) curve (AUC). Potential interaction terms between legally blind status and key covariates were tested, but none showed statistically significant interaction effects. Analyses were conducted using Stata SE 18.5 (College Station, TX, 2023), with statistical significance set at p < 0.05. Small cell sizes (<11) were not reported per HCUP data use agreements.

3.1 Participants

From a total of over 41 million hospitalizations over 2016 to 2021, we identified 338,578 hospitalizations (weighted 1,692,889 hospitalization) for cervical or lumbar spinal fusion that met our inclusion criteria. Of these, 315 (weighted 1575) hospitalizations involved legally blind patients.

3.2 Comorbidities and discharge disposition (Descriptive data)

Compared to the control group, legally blind patients were older (mean age 63 vs. 60 years, p < 0.001), more likely to be on Medicare (71.0 % vs. 42.9 %, p < 0.001), and had higher prevalence of complicated hypertension (18.7 % vs. 7.5 %, p < 0.001), congestive heart failure (7.0 % vs. 3.0 %, p < 0.001), hyperlipidemia (42.9 % vs. 31.7 %, p < 0.001), fall history (6.0 % vs. 2.9 %, p < 0.01), and MMC (11.1 % vs. 5.2 %, p < 0.001) (Table 4). They were also significantly more likely to be discharged to NHD (40.0 % vs. 16.6 %, p < 0.001) and to experience ELOS (50.5 % vs. 32.3 %, p < 0.001) (Table 1).

3.3 Perioperative complications (Outcome data)

After adjusting for age, sex, race, median household income quartile, primary payer, hospital teaching status, obesity, hypertension, congestive heart failure, hyperlipidemia, trauma status, fall history, fragility fracture history, the legally blind group was significantly associated with increased odds of all three adverse outcomes. Legally blind group had 58 % higher odds of experiencing major medical complications (aOR = 1.58, 95 % CI: 1.08–2.32), more than double the odds of being discharged to a non-home setting (aOR = 2.57, 95 % CI: 1.98–3.32), and 67 % higher odds of experiencing a prolonged length of stay (aOR = 1.67, 95 % CI: 1.31–2.12)(Table 2)(Table 3)(Table 4).

4.1 Medical complications

Legally blind patients were significantly more likely to experience the composite outcome of major medical complications following spinal fusion surgery. Some contributing factors to the increased risk of medical complications may include limited mobility due to visual impairment. Previous studies have demonstrated that limitations in early mobilization following spine surgery may increase the risk of cardiopulmonary complications including DVT/PE and pneumonia.^8,9 Patient with visual impairment may be less likely to ambulate independently, face challenges with safe ambulation and rely more heavily on supervised physical therapy sessions for out of bed. While our study could not evaluate these mechanisms directly, our findings suggest that comprehensive ambulation support and structured physical therapy goals may be particularly important in select medical complications in visually impaired patients, especially given their higher baseline fall risk and mobility limitations.

4.2 Extended length of stay and non-home discharge

We also found that legally blind patients had significantly higher odds of extended hospitalization and non-home discharge. These results are consistent with previous studies showing that patients with vision impairment are at greater risk for prolonged length of stay across a variety of medical and surgical indications including joint replacement, pneumonia, sepsis and chronic obstructive pulmonary disease.^10,11 Specifically within the context of orthopedic surgery, a study evaluating perioperative course in legally blind patients undergoing total hip arthroplasty reported similar findings, with legal blindness independently associated with longer length of stay and increased discharge to skilled nursing facilities or rehabilitation centers.¹² Other factors in patients undergoing spinal fusion surgery, including increased age, living alone, higher BMI and higher Oswestry disability index, have also been linked to extended length of stay.¹³ While our dataset could not directly assess physical therapy interventions, previous research has shown that early PT can increase the odds of early discharge, independent of comorbidity profiles.¹⁴ This raises the possibility that tailored rehabilitation strategies could mitigate some of the risks we observed in legally blind patients. Overall, our findings contribute to this existing body of evidence suggesting that visual impairment is a consistent predictor of delays in recovery and increased need for subacute care.

4.3 Clinical implications

Taken together, our findings demonstrate that legal-blindness, and visual impairment more broadly, should be recognized as a distinct risk factor in the context of spinal fusion surgery. Incorporating vision status into preoperative counseling and shared decision making with patients may improve risk communication with patients. Early and intentional involvement of physical therapy, occupational therapy and case management may help mitigate the risk of poor postoperative outcomes and facilitate safe discharge planning. Previous research has shown that multimodal rehabilitation regimens including exercise therapy and cognitive behavioral training may be superior to exercise therapy alone, an approach that could be particularly beneficial in the legally blind population.¹⁵ Care teams should also consider providing enhanced support for patients with visual impairment, including enhanced fall prevention strategies and additional support with ambulation and activities of daily living during the postoperative period.

4.4 Limitations

The most significant limitation of this large, nationally representative inpatient database study was the reliance on administrative coding for classification of legal-blindness. There is a high potential for underreporting of legal-blindness, which could bias prevalence estimates. Additionally, the NIS database is limited to inpatient course, inhibiting our ability to evaluate important postoperative outcomes in spine surgery including pseudoarthrosis and revision procedures. Addressing additional confounders like granular socioeconomic status indicators, social support and access to rehabilitation services is also not possible with the NIS database. Finally, the absolute number of blind patients undergoing spinal fusion was relatively small, limiting the ability to conduct detailed subgroup analyses.