Skip to main content
NCBI home page
American Journal of Speech-Language Pathology logoLink to American Journal of Speech-Language Pathology
. 2024 Aug 1;33(5):2487–2499. doi: 10.1044/2024_AJSLP-23-00264

The Relationship Between Poststroke Dysphagia and Poststroke Depression and Its Risk Factors

Janet Horn a,✉, Kit N Simpson b, Annie N Simpson b, Leonardo F Bonilha c,d, Heather S Bonilha a,e
PMCID: PMC11427738  PMID: 39088240

Abstract

Purpose:

A diagnosis of dysphagia and/or depression after stroke can impact the physical, psychological, and social welfare of stroke survivors. Although poststroke depression (PSD) and poststroke dysphagia are known to occur concurrently, there is a paucity of research that has specifically investigated their association. Therefore, we aimed to study the relationship between PSD and poststroke dysphagia during acute inpatient hospitalization and within 90 days after discharge. Furthermore, we aimed to evaluate the odds and hazard of being diagnosed with depression after stroke and estimate the time to depression diagnosis from the initial stroke diagnosis in patients with and without a diagnosis of dysphagia.

Method:

Using the acute inpatient hospital data set from our previous work, we pulled additional postdischarge administrative claims data from the 2017 Medicare 5% Limited Data Set and conducted a retrospective, cross-sectional study of patients diagnosed with poststroke dysphagia and PSD.

Results:

Patients diagnosed with poststroke dysphagia had 2.7 higher odds of being diagnosed with PSD and had an approximately 1.75-fold higher hazard for PSD diagnosis in the 90 days after discharge compared to patients not diagnosed with dysphagia. Risk factors for PSD included having dysphagia, being female, and having dual eligibility.

Conclusions:

Our results demonstrated a significant association between PSD and poststroke dysphagia. Additional research should further explore the impact of PSD on poststroke dysphagia.

Poststroke dysphagia occurs in up to 78% of stroke survivors (Martino et al., 2005) and is associated with serious complications, negative consequences, and comorbidities, including compromised overall health, decreased quality of life (QOL; Altman et al., 2010; Namasivayam-MacDonald & Shune, 2018), and depressive symptoms (Dziewas et al., 2017; Holland et al., 2011; Verdonschot et al., 2013, 2017). There are also psychosocial effects associated with dysphagia that can negatively impact QOL, such as embarrassment, reduced self-esteem, and social isolation (Ekberg et al., 2002).

Approximately one third of stroke survivors are affected by poststroke depression (PSD), which is an important predictor of poor patient outcomes after stroke (Hackett & Pickles, 2014; Towfighi et al., 2017). PSD is associated with social impairments, reduced QOL, and increased mortality (Bhogal et al., 2004; Bucur & Papagno, 2018; Cole et al., 2001; Das & Rajanikant, 2018; Towfighi et al., 2017). A number of risk factors for PSD have been examined, including demographic, social, and medical factors; however, results across most studies are discordant and, therefore, remain controversial (Babkair, 2017).

The respective impacts of poststroke dysphagia and PSD on the physical, psychological, and social welfare of stroke survivors are well documented (Dziewas et al., 2017; Paolucci et al., 2019); however, other than our previous work establishing the incidence of PSD in patients with poststroke dysphagia (Horn et al., 2022), we found only one other study that had specifically investigated the relationship between depression and poststroke dysphagia (Pritchard et al., 2020). Understanding the link between these co-occurring health conditions is vital for health care providers (i.e., physicians, nurses, and speech-language pathologists) to develop comprehensive treatment plans that address both the physical and psychological aspects of poststroke recovery, ultimately leading to improved patient outcomes and QOL. Thus, to address this important clinical issue and gap in the literature, our study aimed to examine the relationship between poststroke dysphagia and PSD, evaluate the odds and hazard of being diagnosed with depression after stroke, and estimate the time to depression billing diagnosis from the initial stroke diagnosis in patients with and without a diagnosis of dysphagia.

Method

Study Design and Sample

Administrative data from the 2017 Medicare 5% Limited Data Set (LDS) were used to conduct a retrospective, cross-sectional study of individuals with acute ischemic stroke (AIS). This study was not considered human subject research per the university institutional review board.

Description of Administrative Database and Data Set Construction

Medicare is a federal health insurance program in the United States that offers medical coverage predominantly for individuals aged 65 years and older. LDS files are supplied by the Centers for Medicare & Medicaid Services to support research efforts. The 5% LDS files contain de-identified, patient-level claims data derived from a nationally representative 5% random sample of Medicare insurance recipients.

For this study, we followed the same data set construction steps from our original study and used the same acute inpatient hospital data set from that previous work (Horn et al., 2022). We also constructed a postdischarge data set from the same patients used in the acute inpatient hospital data set by reviewing discharge dates and extracting relevant claims data within a 90-day follow-up window after acute hospital discharge. Both the acute inpatient hospital data set and the 90-day postdischarge data set were extracted from 2017 Medicare 5% LDS files.

Data Coding

International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) medical diagnostic codes were utilized for study variables. For depression, we used the ICD-10-CM codes F32.x and F33.x. For dysphagia, we used R13.x, I69.091, I69.191, I69.291, I69.391, I69.891, and I69.991. For (ischemic) stroke, we used I63.x.

Patient Population

The study population consisted of Medicare beneficiaries greater than or equal to 65 years of age with a primary diagnosis of AIS who were also diagnosed with dysphagia and/or depression during their index hospital admission (i.e., initial inpatient hospital admission for a specific condition) and/or within 90 days after discharge. Patients with a diagnosis of prestroke dysphagia or depression within 90 days before stroke were excluded.

Outcome Measures

Our outcomes of interest included diagnosis of dysphagia and/or depression during the AIS acute inpatient hospitalization or within a 90-day follow-up window after discharge, odds and hazard rate of depression diagnosis, and time from stroke to depression diagnosis.

Study Variable Definitions

Clinically relevant variables were defined as follows:

  • Charlson Comorbidity Index (CCI): A validated instrument considered the “gold standard” for predicting the risk of mortality associated with comorbid diseases (Charlson et al., 1987, 2022; Quan et al., 2005, 2011). The index has 19 (original; Charlson et al., 1987) or 17 (modified; Deyo et al., 1992) comorbidity categories, each with an associated weight. A score of 0 signifies the absence of comorbidities, and as the score increases so does the risk of poor outcomes and mortality (Charlson et al., 1987, 2022).

  • Stroke Administrative Severity Index (SASI): A validated measure to quantify stroke severity at the time of hospital discharge designed for application with administrative billing data (Simpson et al., 2018). Index items include aphasia, coma, dysarthria and/or dysphagia, hemiplegia or monoplegia, neglect, nutritional infusion, and tracheostomy and/or ventilation, with scores of 0 indicative of mild, 1–6 indicative of moderate, and 7–31 indicative of severe stroke severity (Simpson et al., 2018).

  • Length of stay (LOS): During a single hospitalization, the time between admission and discharge. Prolonged LOS is a marker of poorer functional status (Chin et al., 2001).

  • Dual eligibility: Describes dually eligible beneficiaries of both Medicare and Medicaid (a U.S. government health insurance program for individuals with limited income), indicative of low socioeconomic status (SES; Moon & Shin, 2006).

  • Tissue plasminogen activator (tPA) administration: Treatment for AIS that dissolves the clot and restores blood flow to areas of the brain impacted by stroke (National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group, 1995), indicative of higher quality of care and improved clinical outcomes (Saver et al., 2013).

Statistical Analysis

Statistical analyses were performed using SAS statistical software (Version 9.4, released 2016, SAS Institute, Inc.). Demographics and clinical characteristics were compared using t tests for continuous variables and chi-square for categorical variables; p values were considered statistically significant for α < .05. For regression analyses, unadjusted results demonstrated a bivariate relationship between each predictor and the outcome, while adjusted results demonstrated potential interactions and confounding effects of multiple variables simultaneously. In our adjusted results, we controlled for covariates, whereas in our unadjusted results, we did not control for covariates.

Regression coefficients and adjusted odds ratios (ORs) were estimated using logistic regression (Proc Logistic in SAS) to determine if PSD is associated with poststroke dysphagia and if patients diagnosed with poststroke dysphagia were more likely to be diagnosed with PSD. Using purposeful selection, clinically relevant predictor variables were included (Hosmer et al., 2013; Stoltzfus, 2011) in the initial model, such as age (in years), gender (male/female), race (White, Black, Hispanic, or other), CCI (score), SASI (score), tPA (yes/no), LOS (in days), and dual eligibility (yes/no). For model building, a direct approach was used, simultaneously placing all predictor variables identified for inclusion with equal importance into a multivariable model (Hosmer et al., 2013; Stoltzfus, 2011). Multicollinearity between predictor variables was checked by examining several values, including variables with high correlation (> .8 indicating multicollinearity), a variance inflation factor greater than 10, tolerance values greater than .1, and small Eigenvalues (close to 0) with large corresponding condition values (indicating multicollinearity; Schreiber-Gregory, 2017). Given that multicollinearity was not detected, model building was continued using a less stringent variable inclusion criterion (α < .25) so as not to exclude potentially important variables at this initial stage of model development (Hosmer et al., 2013; Mickey & Greenland, 1989; Stoltzfus, 2011). Variables that were not significant at p < .25 were manually removed one at a time, and the model was refit using the traditional level of statistical significance (p < .05) until a parsimonious model was constructed.

Time-to-event (survival) analyses (including the Cox proportional hazards model) were conducted to assess the time to depression after stroke and estimate the hazard (hazard ratio [HR]) for being diagnosed with PSD in patients with and without a dysphagia diagnosis. Time-to-event analysis (Proc Lifetest in SAS) was used to estimate the unadjusted time to diagnosis of depression (event) from the initial diagnosis of stroke (in days) by dysphagic versus nondysphagic groups without controlling for covariates. Then conventional Cox proportional hazards models (Proc Phreg in SAS) were constructed to determine which covariates (age, gender, race, CCI, SASI, tPA, LOS, or dual eligibility) were significantly associated with the time to PSD diagnosis using adjusted HRs. Next, each covariate was manually removed if it did not meet inclusion criteria defined as adequate model fit statistic, likelihood ratio tests, and statistical significance (< .05). We tested for interaction effects between covariates, and lastly, we included all significant covariates in the final parsimonious model. After the final model was constructed, we performed diagnostics to check for adequacy of the model.

A fundamental assumption of Cox regression is that hazards between groups are constant (or proportional) over time (Bellera et al., 2010; Schober & Vetter, 2018; UCLA: Statistical Consulting Group, n.d.). If this assumption of proportionality is violated, biased and/or incorrect estimates may be derived, resulting in misleading interpretations (Bellera et al., 2010; UCLA: Statistical Consulting Group, n.d.). Thus, we assessed the proportionality of the hazards using graphical checks for categorical covariates by which Kaplan–Meier survival curves are plotted for each level of categorical covariate, and then the survival function graphs were judged as to whether or not the survival curves appear parallel (with a parallel graph indicative of proportionality; Bellera et al., 2010; Fisher & Lin, 1999; UCLA: Statistical Consulting Group, n.d.). Typically, graphical checks alone are not sufficient to assess proportionality due to their subjectivity (Bellera et al., 2010); however, it was evident that survival function graphs for all categorical covariates we assessed (dysphagia, gender [female], race [White], dual eligibility) were not parallel (displayed crossed curves). This was suggestive of nonproportionality, meaning there was an interaction between these covariates and time (Bellera et al., 2010; UCLA: Statistical Consulting Group, n.d.). For the continuous covariate (age), we applied the empirical score process using a transform of the martingale residuals as a diagnostic for proportionality (Lin et al., 1993). Then, we inspected the simulation graph for an aberrant observed pattern and checked the corresponding supremum test results for significance (p < .05), indicative of a violation of the proportional hazards assumption (Allison, 2010; Lin et al., 1993; UCLA: Statistical Consulting Group, n.d.). No violation was detected for the continuous covariate (age).

To account for the nonproportionality of four of the covariates (dysphagia, female, White, dual eligibility), we created time-dependent variables that explicitly introduced Covariate × Time interactions into the Cox model, which generalizes the model to permit the use of nonproportional hazards, thereby addressing the proportionality violation (Allison, 2010; Bellera et al., 2010; Cox, 1972; UCLA: Statistical Consulting Group, n.d.). After running the Cox models again with each Covariate × Time interaction term, we found that the interaction covariates for dysphagia and White remained significant, indicating nonproportionality; however, our use of the method for extending the Cox model by including Covariate × Time interactions as predictors allowed for the incorporation of nonproportionality in the Cox model (Allison, 2010; UCLA: Statistical Consulting Group, n.d.).

Results

In a data set of 9,163 patients (Horn et al., 2022), the mean age was 78.66 years (SD = 8.56), and age range was 65–98 years. Fifty-three percent of patients were women, and 82% of patients were White. (Patient characteristics and descriptive data are described in Table 1.) There were 1,440 (15.7%) individuals diagnosed with dysphagia during inpatient hospitalization. While there were no significant differences in gender or race between patients with and without a diagnosis of dysphagia, those with dysphagia were significantly older and had higher CCI and SASI scores. Dysphagic individuals also experienced greater hospital stays, with a mean LOS of 7.99 days (SD = 5.76) compared to 4.83 days (SD = 3.69) for those without dysphagia (p < .0001). In addition, those diagnosed with dysphagia demonstrated significantly higher rates of depression diagnosis than those not diagnosed with dysphagia during acute hospitalization, 12.0% versus 9.5%, respectively (p = .003).

Table 1.

Baseline characteristics and descriptive data.

Characteristics General stroke
N = 9,163 (100%)		 Dysphagia
n = 1,440 (15.72%)		 No dysphagia
n = 7,723 (84.28%)		 p
Age, yearsa 78.66 (8.56)
65–98		 80.46 (8.76)
65–98		 78.33 (8.48)
65–98		 < .0001
CCI, max possible 24a 3.82 (2.18)
1–17		 4.35 (2.19)
1–15		 3.72 (2.16)
1–17		 < .0001
SASI, max possible 56a 6.04 (6.84)
0–48		 9.31 (6.60)
0–45		 5.44 (6.71)
0–48		 < .0001
LOS, daysa 5.33 (4.24)
1–93		 7.99 (5.76)
1–56		 4.83 (3.69)
1–93		 < .0001
Femaleb 4,901 (53.49) 780 (54.17) 4,121 (53.36) .573
Race .115
 Whiteb 7,532 (82.20) 1,157 (80.35) 6,375 (82.55)
 Blackb 1,042 (11.37) 171 (11.88) 871 (11.28)
 Hispanicb 132 (1.44) 25 (1.74) 107 (1.39)
 Otherb 457 (4.99) 87 (6.04) 370 (4.79)
Depressionb 908 (9.91) 173 (12.01) 735 (9.52) .003
Cognitive declineb 1,883 (20.55) 421 (29.24) 1,462 (18.93) < .0001
tPAb 460 (5.02) 50 (3.47) 410 (5.31) .003
Dual eligibilityb 1,616 (17.64) 292 (20.28) 1,324 (17.14) .004
Open in a new tab

Note. Data are from Horn et al. (2022). CCI = Charlson Comorbidity Index; SASI = Stroke Administrative Severity Index; LOS = length of stay; tPA = tissue plasminogen activator.

a

Values are mean (SD) and range.

b

Values are n (%).

Odds of PSD

Multivariable logistic regression was used to examine the relationship between PSD and poststroke dysphagia. All predictor variables were included in the initial model (see Table 2). After nonsignificant variables were manually removed and collinearity was assessed, the final parsimonious model was run (see Table 3). The Hosmer–Lemeshow (HL) goodness-of-fit (GOF) test of a random 10% sample of the population (n = 917) demonstrated that the model was a good fit with an insignificant p value of > .05 (p = .7984).

Table 2.

Initial multivariable logistic regression model to determine association between poststroke depression and poststroke dysphagia with all potential covariates included.

Variable Coefficient SE OR 95% CI p
Poststroke dysphagia 0.9904 0.0828 2.692 [2.289, 3.167] < .0001
Age −0.2714 0.0441 0.762 [0.699, 0.831] < .0001
Female 0.5438 0.0759 1.722 [1.484, 1.999] < .0001
Racea
 White 0.5722 0.1421 1.850 [1.435, 2.384] < .0001
 Hispanic −0.7632 0.3858 0.487 [0.174, 1.358] .048
 Other 0.2339 0.1815 1.319 [0.884, 1.967] .198
CCI −0.00407 0.0172 0.996 [0.963, 1.030] .813
SASI −0.00876 0.00573 0.991 [0.980, 1.002] .126
tPA 0.2426 0.1546 1.275 [0.941, 1.726] .117
LOS 0.0145 0.00783 1.015 [0.999, 1.030] .065
Dual eligibility 0.3497 0.0917 1.419 [1.185, 1.698] .0001
Intercept −1.2083 0.3661 — — .001
Open in a new tab

Note. SE = standard error; OR = odds ratio; CI = confidence interval; CCI = Charlson Comorbidity Index; SASI = Stroke Administrative Severity Index; tPA = Tissue plasminogen activator; LOS = length of stay in acute hospital.

a

Reference group was Black patients.

Table 3.

Final multivariable logistic regression model with significant covariates included and the Hosmer–Lemeshow goodness-of-fit test.

Variable Coefficient SE OR 95% CI p
Poststroke dysphagia 0.9972 0.0804 2.711 [2.315, 3.174] < .0001
Age −0.2714 0.0441 0.762 [0.699, 0.831] < .0001
Female 0.5383 0.0757 1.713 [1.477, 1.987] < .0001
Racea
 White 0.5684 0.1419 1.818 [1.413, 2.340] < .0001
 Hispanic −0.7677 0.3857 0.478 [0.171, 1.333] .047
 Other 0.2289 0.1814 1.295 [0.868, 1.931] .207
Dual eligibility 0.3553 0.0912 1.427 [1.193, 1.706] < .0001
Intercept −1.1812 0.3587 .001
n % of population χ 2 df p
Hosmer–Lemeshowb 917 10 4.6094 8 .7984
Open in a new tab

Note.  SE = standard error; OR = odds ratio; CI = confidence interval.

a

Reference group was Black patients.

b

Hosmer–Lemeshow goodness-of-fit test applied to a 10% random sample of the population.

We conducted multivariable logistic regression analysis on the entire population (N = 9,163), which revealed that patients who were diagnosed with poststroke dysphagia were 2.7 times more likely to have a diagnosis of depression within 90 days after discharge from the acute hospital (adjusted OR = 2.711, 95% CI [2.315, 3.174], p < .0001) compared to patients who were not diagnosed with poststroke dysphagia. White patients were 81.8% more likely to be diagnosed with PSD than Black patients (adjusted OR = 1.818, 95% CI [1.413, 2.340], p < .0001). The odds of being diagnosed with PSD was higher in women with poststroke dysphagia by 71.3% (adjusted OR = 1.713, 95% CI [1.477, 1.987], p < .0001) and in individuals who qualified for dual eligibility by 42.7% (adjusted OR = 1.427, 95% CI [1.193, 1.706], p < .0001). The odds of depression diagnosis decreased by 23.8% with each added year of age (adjusted OR = 0.762, 95% CI [0.699, 0.831], p < .0001).

Unadjusted Time to Depression

Without covariate adjustment, the unadjusted estimation of mean time from diagnosis of AIS to diagnosis of depression was 40 days (SD = 30.69) for patients diagnosed with dysphagia and 34 days (SD = 30.18) for patients not diagnosed with dysphagia. Although patients who had a diagnosis of poststroke dysphagia demonstrated greater mean days to depression diagnosis than patients who did not have a diagnosis of poststroke dysphagia, these results were not significant (p = .729). The unadjusted survival plot (see Figure 1) illustrates the differences in time to depression diagnosis between those who were diagnosed with dysphagia and those who were not diagnosed with dysphagia. It showed that both groups had an equally high probability of survival (not being diagnosed with depression) soon after discharge, and as the 90-day postdischarge period progressed, patients without a dysphagia diagnosis demonstrated better survivability (less likelihood of being diagnosed with depression) than patients with a dysphagia diagnosis. Furthermore, patients who were diagnosed with dysphagia appeared to experience a steady increase in diagnosis of PSD compared to those not diagnosed with dysphagia over the 90-day postdischarge period, with the dysphagic group demonstrating an approximately 1.5-fold higher probability of being diagnosed with PSD during the last 15 days of follow-up.

Figure 1.

A plot of the survival probability on the y axis and the time in days on the x axis. The curve for persons with dysphagia passes through (0, 1) and (90, 0.85). The curve for persons without dysphagia passes through (0, 1) and (90, 0.95).

Open in a new tab

Survival curve for poststroke depression during 90-day postdischarge follow-up period by the presence of dysphagia diagnosis, Kaplan–Meier method.

Unadjusted initial Cox proportional hazards models were run on each covariate separately to determine potential influence on time to depression. We found a significant effect for dysphagia, age, gender (female), race (White), LOS, and dual eligibility. Unadjusted Cox models showed that the hazard for depression diagnosis in patients who have been diagnosed with poststroke dysphagia was 2.4-fold greater (HR = 2.420, 95% CI [2.099, 2.790], p < .0001) than the hazard of depression diagnosis in those without a dysphagia diagnosis. Conversely, with each increase in year of age after discharge from the hospital, the hazard for diagnosis of PSD decreased by 1.5% (HR = 0.985, 95% CI [0.978, 0.993], p = .0002). The hazard for women was approximately 54% (HR = 1.541, 95% CI [1.345, 1.766], p < .0001) greater than the hazard for men, while the hazard of depression diagnosis for White patients was 37% (HR = 1.37, 95% CI [1.087, 1.727], p = .0077) greater than the hazard for depression diagnosis for non-White patients. For each 1-day increase in LOS, the hazard of PSD diagnosis increased by 2.2% (HR = 1.022, 95% CI [1.010, 1.034], p = .0002). The hazard of PSD diagnosis for those who qualified for dual eligibility was about 41% (HR = 1.414, 95% CI [1.210, 1.652], p < .0001) greater than the hazard for those who did not qualify for dual eligibility. No significant effects were observed for Hispanic or “other” race, CCI, SASI, or tPA.

Adjusted Time to Depression

Results from the extended Cox model with covariate adjustment (see Table 4) were comparable to unadjusted results. Based on the model, having a diagnosis of dysphagia, being female, being White, and having dual eligibility significantly increased the risk of being diagnosed with depression after stroke, while every year of age significantly decreased the risk of depression diagnosis. The Dysphagia × Time interaction covariate that we incorporated into the extended Cox model allowed the effect of dysphagia (our covariate of greatest interest) to change with time (Allison, 2010), and its significance suggested that the dysphagia effect did, in fact, vary over time since discharge from the hospital. This is illustrated by the adjusted cumulative hazard plot (see Figure 2), which showed that when controlling for significant covariates, the hazard for diagnosis of PSD was initially low (< .01) until about the 10th day after discharge and then consistently increased throughout the duration of the 90-day follow-up period for both patients with and without a diagnosis of poststroke dysphagia. On average and on any given day in the 90 days after discharge, the hazard for diagnosis of depression for patients who had a dysphagia diagnosis was approximately 76% greater (HR = 1.755, 95% CI [1.368, 2.251], p < .0001) than the hazard for patients who did not have a dysphagia diagnosis. In addition, the hazard of PSD diagnosis for women was about 67% higher (HR = 1.666, 95% CI [1.449, 1.915], p < .0001) than the hazard for men, while the hazard of depression diagnosis for individuals with dual eligibility was approximately 40% higher (HR = 1.404, 95% CI [1.193, 1.654], p < .0001) than the hazard for those who did not have dual eligibility. In contrast, hazard decreased significantly by 2.5% (HR = 0.975, 95% CI [0.967, 0.982], p < .0001) with each year of age, suggesting that older stroke survivors are less likely to be diagnosed with depression. When controlling for significant covariates, we found that the hazard of PSD diagnosis for White patients was 71% higher (HR = 1.708, 95% CI [1.401, 2.082], p < .0001) than the hazard for non-White patients. We ran additional Cox models to determine if there were any significant interaction effects between dysphagia and other significant covariates (age, gender [female], race [White], LOS, dual eligibility) for time to depression and found none.

Table 4.

Final parsimonious Cox proportional hazards model with significant covariates.

Variable Coefficient SE HR 95% CI p
Poststroke dysphagia 0.56246 0.56246 1.755 [1.368, 2.251] < .0001
Female 0.51049 0.07109 1.666 [1.449, 1.915] < .0001
Race, White 0.53525 0.10101 1.708 [1.401, 2.082] < .0001
Dual eligibility 0.33958 0.08341 1.404 [1.193, 1.654] < .0001
Age (years) −0.02581 0.00412 0.975 [0.967, 0.982] < .0001
Open in a new tab

Note. SE = standard error; HR = hazard ratio; CI = confidence interval.

Figure 2.

A plot of the cumulative hazard on the y axis and the time in days on the x axis. The curve for persons with dysphagia passes through (5, 0), (60, 0.07), and (90, 0.085). The curve for persons without dysphagia passes through (5, 0), (60, 0.06), and (90, 0.078).

Open in a new tab

Hazard curve for poststroke depression during 90-day postdischarge follow-up period by the presence of dysphagia diagnosis with covariates.

The adjusted cumulative hazard plot (see Figure 2) showed that when controlling for significant covariates, there was a low immediate hazard for depression diagnosis followed by a steady increase in hazard for both groups over time, continuing until the end of the follow-up period. The group with a diagnosis of dysphagia demonstrated a higher hazard for depression diagnosis almost immediately after discharge and for the duration of the 90-day follow-up period compared to the group without a diagnosis of dysphagia. Furthermore, those diagnosed with poststroke dysphagia demonstrated an approximately 1.75-fold higher hazard for PSD diagnosis than those not diagnosed with poststroke dysphagia, suggesting that at any given time in the 90 days after discharge, the hazard for depression diagnosis is higher for patients diagnosed with dysphagia compared to patients not diagnosed with dysphagia.

Discussion

In this study of 9,163 patients with AIS, we found that depression diagnosis was significantly affected by dysphagia diagnosis during the 90-day postdischarge follow-up period. Those who had been diagnosed with dysphagia were almost 3 times more likely to be diagnosed with depression compared to patients who had not been diagnosed with dysphagia. In addition, patients diagnosed with dysphagia demonstrated an approximately 1.75-fold higher hazard for PSD diagnosis than patients not diagnosed with dysphagia.

Our findings provide evidence in support of a relationship between poststroke dysphagia and PSD. To our knowledge, the only other study to specifically examine depression in patients with stroke-induced dysphagia as a primary outcome was conducted by Pritchard et al. (2020). Pritchard et al. explored the association between stroke-induced dysphagia and depression during the subacute stage of recovery in their retrospective, cross-sectional study. They used the Stroke Recovery in Underserved Populations 2005–2006 database to investigate whether poststroke dysphagia during an inpatient rehabilitation facility (IRF) stay was a significant predictor of depressive symptoms at a 3-month follow-up (Pritchard et al., 2020). Depressive symptoms were measured using the Center for Epidemiologic Studies Depression Scale with a total score of > 16 (the cutoff) classified as depressed (Cosco et al., 2020; Sheehan et al., 1995). The investigators found that at the 3-month follow-up, individuals with poststroke dysphagia during their IRF stay demonstrated a greater likelihood of developing depressive symptoms compared to those without dysphagia (Pritchard et al., 2020). Pritchard et al.'s results are consistent with the findings of this study, that having a diagnosis of dysphagia after stroke was positively associated with also having a diagnosis of depression.

Odds of PSD Diagnosis

In the present study, we found that having a diagnosis of dysphagia, being female, being White, and having dual eligibility were positively associated with depression diagnosis after stroke, while age was negatively associated with PSD diagnosis. Although there were very few studies by which to compare our findings, we found that our results were similar to Ayerbe et al.'s (2011) findings that at 3 months poststroke, dysphagic patients were almost twice as likely to have depression than nondysphagic patients; however, Saxena et al. (2008) found no significant association between PSD and dysphagia at 6 months after stroke (in multivariate analyses). The reason for the disparity may be due to variations in study methodology, sample sizes, diagnostic assessments, and time frames. Despite the discordance with Saxena et al.'s results, our findings are supported by other nonstroke dysphagia literature, which posit that dysphagia is associated with affective complaints (Verdonschot et al., 2013, 2017). In a systematic review by Verdonschot et al. (2017), the researchers found that all 24 articles they appraised suggested that depressive symptoms were significantly and positively associated with dysphagia. Furthermore, dysphagia is known to have severe social and psychological impacts that reduce QOL, such as embarrassment, isolation, and loss of self-esteem due to swallowing difficulties and anxiety, panic, and avoidance of meals due to fear of food sticking in the throat and choking (Ekberg et al., 2002). For these reasons, it is plausible that dysphagia is not only associated with PSD but also related to increased odds of PSD in poststroke patients, as we found in the present study.

Sienkiewicz-Jarosz et al.'s (2010) results revealed that stroke patients with more depressive symptoms had lower income, which was comparable to our findings with patients who had dual eligibility (an indicator of low SES), as were Khedr et al.'s (2021) results that PSD is significantly associated with lower SES. It is known that low SES is a determinant of poor health status, and the combination of low SES and high levels of psychological distress has been shown to have a multiplicative effect in which low SES magnifies the negative effects of psychological distress (Lazzarino et al., 2013a, 2013b). Furthermore, it is believed that those with lower SES possess fewer financial, social, and psychological resources to manage adverse events (Lazzarino et al., 2013a, 2013b; Matthews & Gallo, 2011); therefore, our findings that individuals with dual eligibility have increased odds of PSD compared to those who do not have dual eligibility are in line with the literature.

With regard to gender differences, we found that women were more likely to be diagnosed with PSD than men, which was similar to Goldmann et al.'s (2016) findings that the odds of PSD were higher in women. Conversely, our results contradicted Ayasrah et al.'s (2018) findings in which gender was not found to predict PSD. In the 15 studies examining prevalence of depression that Appelros et al. (2010) reviewed, most reported that PSD was more common in women than men, while others found no difference between women and men, and one study found a higher prevalence of PSD in men. The authors concluded that whether or not being a woman is a risk factor for PSD remains ill-defined (Appelros et al., 2010). Even so, we know that there are gender-related differences in depression outside the setting of stroke, with women almost twice as likely than men to become depressed during their lifetime (Kuehner, 2017; Salk et al., 2017). The reasons for this “gender gap” in depression remain unclear; however, there is evidence to support several possible factors, including genetic susceptibility, hormonal influences, psychological stress responsiveness, coping styles, and social roles (Kuehner, 2017). For these reasons, it is plausible that after stroke, women are more likely to be diagnosed with depression than men, as we found in the present study.

We found the odds of PSD diagnosis decreased with age, which is in agreement with Goldmann et al.'s (2016) findings that older patients demonstrated lower odds of PSD; however, our findings differed from Ayasrah et al.'s (2018) results in which age was not found to predict PSD. McCarthy et al. (2016) reported that at 3 months poststroke, patients between 25 and 64 years of age had significantly greater depressive symptoms than those 65 years and older, with the 25–54 years age group demonstrating the highest risk for depression out of all age groups examined, which is in line with our findings that older patients have lower odds of PSD diagnosis. The authors noted, however, that their results were in accordance with some similar studies but contradicted the results of others and suggested that these disparities were due to variability in study methods, such as using age as a dichotomous, instead of continuous, variable, and restricting age groups in study samples (McCarthy et al., 2016). Furthermore, McCarthy et al. proposed that the association between age and PSD may be curvilinear as opposed to linear, meaning that risk of PSD is greatest between the ages of 25 and 54 years, attenuating through midlife and early old age and then increasing again in late old age. Another potential reason for the inconsistent findings in studies examining age and PSD could be the underdiagnosis of depression in older individuals as a result of age-related clinician bias in which depression is assumed to be a normal response to a serious medical event, disease diagnosis, or even advancing age (Stewart, 2004). Clinicians may observe symptoms of depression but not diagnose or refer an older patient for intervention if these symptoms are assumed to be expected in this particular patient age group.

While we found that White patients were more likely to be diagnosed with PSD than any other race within 3 months after discharge from the acute hospital, it should be noted that we had a disproportionately large number of White patients in our population (> 80%). Given this, we are reticent to make conclusions or form hypotheses based on this result. Prior literature, all with various study design and method limitations, is equivocal. Thus, there is support for and against racial differences in the rate of PSD diagnosis.

Hazard of PSD Diagnosis

We are unable to directly compare our results to analogous studies because, to our knowledge, there are no studies that have examined the hazard of PSD diagnosis in patients diagnosed with poststroke dysphagia. Therefore, our findings may be the first to establish that diagnosis of dysphagia has a highly significant effect on the hazard of diagnosis of depression after AIS during the first 90 days after discharge from the hospital. Nevertheless, we can compare our findings to the few published studies that have assessed hazard. For example, Aben et al. (2003) included female gender and age in their Cox regression analyses to assess the effect of cohort (stroke vs. myocardial infarction) on the cumulative incidence of depression, and though they did not discuss the hazard of these variables explicitly, their results showed that, similar to our findings, the hazard of depression in the stroke cohort was significantly higher (about 60%) in women versus men. Contrary to our findings that with every year of age the hazard of PSD diagnosis significantly decreased, age was not found to be a significant covariate in Aben et al.'s study. Likewise, Leentjens et al.'s (2006) findings that female gender was not significant contradicted our results that being female significantly increased the hazard of being diagnosed with depression after stroke. As with studies reporting the odds of PSD diagnosis, our findings are consistent with some results from studies examining hazard, yet contrary to others. As previously discussed, the reasons for these differences may be variations in methodology, sample size, inclusion/exclusion criteria, diagnostic timing and tools, and so on.

Time to Depression Diagnosis

With regard to time to depression diagnosis, to our knowledge, there have been no previous studies that have examined time from initial stroke diagnosis to the first time depression diagnosis appears in billing data during an index time period in patients diagnosed with poststroke dysphagia; thus, the present study appears to be the first on this specific topic.

Later Onset of PSD Versus Later Diagnosis of PSD

One reason for our findings that patients with a diagnosis of dysphagia were diagnosed with depression later than those without a diagnosis of dysphagia could be that depression actually developed during the subacute, as opposed to the acute, phase after stroke, which is feasible given that the literature suggests the timing of PSD is variable among individuals and initial onset can occur between several days to years after stroke (Conroy et al., 2020). Most studies report the highest rates of PSD within the first month to a year after stroke, with a decline in the subsequent 12–24 months thereafter (Conroy et al., 2020; Ostir et al., 2011; Towfighi et al., 2017). Additionally, patients with a diagnosis of poststroke dysphagia often have higher stroke severity (Arnold et al., 2016) and greater functional limitations (Castagna et al., 2019) than those not diagnosed with poststroke dysphagia, requiring continuation of care at an IRF or a skilled nursing facility after discharge. As a result, these patients could potentially develop later-onset PSD when transitioning from a medical facility, where they received 24-hr care, to home, where they experience an abrupt discontinuation of constant care and reduced socialization.

Conversely, it is possible that PSD was present in patients diagnosed with dysphagia earlier than detected. The diagnosis could have been delayed due to underdiagnosis of depression as a result of barriers assessing patients with concurrent cognitive and/or language disorders and misclassification of depressive symptoms as stroke symptoms. Because PSD is associated with worse functional outcomes, increased disability, and higher mortality after stroke (Towfighi et al., 2017; Zhang et al., 2020), the timing of depression diagnosis is important for early identification and prompt intervention; however, optimal screening time after stroke is not known (Towfighi et al., 2017). Nevertheless, it is plausible that early and recurrent depression screenings starting during acute hospitalization would benefit those at risk for PSD.

Limitations

As with our previous study (Horn et al., 2022), there are disadvantages in using large administrative data sets, such as lack of important patient details including premorbid function, diagnostic methods, and disease/disorder severity. For example, our 90-day prestroke dysphagia or depression cutoff does not guarantee that patients included in our data set did not have undiagnosed dysphagia or depression during this time. Conversely, patients who received a diagnosis of depression after the 90-day postdischarge period were not captured in this study. Another major limitation of using administrative billing data is that the sequence of dysphagia and depression onset cannot be definitively determined; therefore, our conclusions are derived from the results of understanding the concomitant illnesses of poststroke dysphagia and PSD versus positing causality. Sampling bias limits the generalizability of results to Medicare beneficiaries 65 years of age and older, excluding younger stroke survivors and individuals without Medicare. Additionally, study findings may be underestimated due to coding errors and misclassification bias, which are known limitations of administrative data (Cohen et al., 2020). Finally, a criticism of the HL GOF test is that it has low power (Stoltzfus, 2011); however, in our case, low power is not a concern because of our large data set (N = 9,163). The concern would be that because power increases as sample size increases, small deviations from the model in a large data set will appear significant (Paul et al., 2013). We have addressed this limitation by conducting the HL GOF test with a smaller 10% random sample of our data set (n = 917).

Clinical Implications

The foundational knowledge gained from this study is a starting point for understanding the factors associated with a diagnosis of PSD in patients with poststroke dysphagia in pursuit of improved identification, earlier intervention, and treatment strategy guidelines to maximize recovery and improve outcomes for this patient population. Furthermore, these results, which establish an association between PSD and poststroke dysphagia, underscore the importance of active and continued monitoring for depressive symptoms in this patient population and the need for implementation of adjunctive screening for depressive symptoms along with poststroke dysphagia screening and/or assessment.

Conclusions

This research has contributed to the existing dysphagia body of knowledge and addressed a gap in the literature by examining a rarely studied population, poststroke dysphagic patients with depression, and revealing new insights into the relationship between poststroke dysphagia and PSD. Results of this study demonstrated evidence of an association between PSD and poststroke dysphagia. That is, patients with poststroke dysphagia had significantly greater odds and hazard of being diagnosed with PSD within 90 days after discharge compared to those without dysphagia. Results revealed that women and individuals with dual eligibility (an indicator of low SES) demonstrated significantly greater odds and hazard for PSD after stroke as well. Furthermore, we found that those diagnosed with dysphagia had higher hazard for PSD diagnosis almost immediately after discharge and throughout the 90-day follow-up period compared to those not diagnosed with dysphagia; therefore, depression screenings in patients diagnosed with poststroke dysphagia by the interdisciplinary dysphagia management team are crucial for identifying these high-risk patients and providing timely and appropriate referrals.

Data Availability Statement

The data sets generated and analyzed during the current study are not publicly available due to license restrictions.

Acknowledgments

This research was supported by the National Institute on Deafness and Other Communication Disorders, National Institutes of Health, Interdisciplinary Research Training in Otolaryngology and Communication Science, through Grant T32 DC014435 to Judy R. Dubno.

Funding Statement

This research was supported by the National Institute on Deafness and Other Communication Disorders, National Institutes of Health, Interdisciplinary Research Training in Otolaryngology and Communication Science, through Grant T32 DC014435 to Judy R. Dubno.

References

  1. Aben, I., Verhey, F., Strik, J., Lousberg, R., Lodder, J., & Honig, A. (2003). A comparative study into the one year cumulative incidence of depression after stroke and myocardial infarction. Journal of Neurology, Neurosurgery, and Psychiatry, 74(5), 581–585. 10.1136/jnnp.74.5.581 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Allison, P. D. (2010). Survival analysis using SAS: A practical guide (2nd ed.). SAS Institute. [Google Scholar]
  3. Altman, K. W., Yu, G. P., & Schaefer, S. D. (2010). Consequence of dysphagia in the hospitalized patient: Impact on prognosis and hospital resources. Archives of Otolaryngology—Head & Neck Surgery, 136(8), 784–789. 10.1001/archoto.2010.129 [DOI] [PubMed] [Google Scholar]
  4. Appelros, P., Stegmayr, B., & Terént, A. (2010). A review on sex differences in stroke treatment and outcome. Acta Neurologica Scandinavica, 121(6), 359–369. 10.1111/j.1600-0404.2009.01258.x [DOI] [PubMed] [Google Scholar]
  5. Arnold, M., Liesirova, K., Broeg-Morvay, A., Meisterernst, J., Schlager, M., Mono, M. L., El-Koussy, M., Kägi, G., Jung, S., & Sarikaya, H. (2016) Dysphagia in acute stroke: Incidence, burden and impact on clinical outcome. PLOS ONE, 11(2), Article e0148424. 10.1371/journal.pone.0148424 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Ayasrah, S. M., Ahmad, M. M., & Basheti, I. A. (2018). Post-stroke depression in Jordan: Prevalence correlates and predictors. Journal of Stroke and Cerebrovascular Diseases, 27(5), 1134–1142. 10.1016/j.jstrokecerebrovasdis.2017.11.027 [DOI] [PubMed] [Google Scholar]
  7. Ayerbe, L., Ayis, S., Rudd, A. G., Heuschmann, P. U., & Wolfe, C. D. (2011). Natural history, predictors, and associations of depression 5 years after stroke: The South London Stroke Register. Stroke, 42(7), 1907–1911. 10.1161/strokeaha.110.605808 [DOI] [PubMed] [Google Scholar]
  8. Babkair, L. A. (2017). Risk factors for poststroke depression: An integrative review. Journal of Neuroscience Nursing, 49(2), 73–84. 10.1097/jnn.0000000000000271 [DOI] [PubMed] [Google Scholar]
  9. Bellera, C. A., MacGrogan, G., Debled, M., Tunon de Lara, C., Brouste, V., & Mathoulin-Pélissier, S. (2010). Variables with time-varying effects and the Cox model: Some statistical concepts illustrated with a prognostic factor study in breast cancer. BMC Medical Research Methodology, 10(1), Article 20. 10.1186/1471-2288-10-20 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Bhogal, S. K., Teasell, R., Foley, N., & Speechley, M. (2004). Lesion location and poststroke depression: Systematic review of the methodological limitations in the literature. Stroke, 35(3), 794–802. 10.1161/01.STR.0000117237.98749.26 [DOI] [PubMed] [Google Scholar]
  11. Bucur, M., & Papagno, C. (2018). A systematic review of noninvasive brain stimulation for post-stroke depression. Journal of Affective Disorders, 238, 69–78. 10.1016/j.jad.2018.05.026 [DOI] [PubMed] [Google Scholar]
  12. Castagna, A., Ferrara, L., Asnaghi, E., Rega, V., & Fiorini, G. (2019). Functional limitations and cognitive impairment predict the outcome of dysphagia in older patients after an acute neurologic event. NeuroRehabilitation, 44(3), 413–418. 10.3233/nre-182635 [DOI] [PubMed] [Google Scholar]
  13. Charlson, M. E., Carrozzino, D., Guidi, J., & Patierno, C. (2022). Charlson Comorbidity Index: A critical review of clinimetric properties. Psychotherapy and Psychosomatics, 91(1), 8–35. 10.1159/000521288 [DOI] [PubMed] [Google Scholar]
  14. Charlson, M. E., Pompei, P., Ales, K. L., & MacKenzie, C. R. (1987). A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. Journal of Chronic Diseases, 40(5), 373–383. 10.1016/0021-9681(87)90171-8 [DOI] [PubMed] [Google Scholar]
  15. Chin, J. J., Sahadevan, S., Tan, C. Y., Ho, S. C., & Choo, P. W. (2001). Critical role of functional decline in delayed discharge from an acute geriatric unit. Annals of the Academy of Medicine, Singapore, 30(6), 593–599. [PubMed] [Google Scholar]
  16. Cohen, S. M., Lekan, D., Risoli, T., Jr., Lee, H. J., Misono, S., Whitson, H. E., & Raman, S. (2020) Association between dysphagia and inpatient outcomes across frailty level among patients ≥ 50 years of age. Dysphagia, 35(5), 787–797. 10.1007/s00455-019-10084-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Cole, M. G., Elie, L. M., McCusker, J., Bellavance, F., & Mansour, A. (2001). Feasibility and effectiveness of treatments for post-stroke depression in elderly inpatients: Systematic review. Journal of Geriatric Psychiatry and Neurology, 14(1), 37–41. 10.1177/089198870101400109 [DOI] [PubMed] [Google Scholar]
  18. Conroy, S. K., Brownlowe, K. B., & McAllister, T. W. (2020). Depression comorbid with stroke, traumatic brain injury, Parkinson's disease, and multiple sclerosis: Diagnosis and treatment. Focus, 18(2), 150–161. 10.1176/appi.focus.20200004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Cosco, T. D., Lachance, C. C., Blodgett, J. M., Stubbs, B., Co, M., Veronese, N., Wu, Y. T., & Prina, A. M. (2020). Latent structure of the Centre for Epidemiologic Studies Depression Scale (CES-D) in older adult populations: A systematic review. Aging & Mental Health, 24(5), 700–704. 10.1080/13607863.2019.1566434 [DOI] [PubMed] [Google Scholar]
  20. Cox, D. R. (1972). Regression models and life-tables. Journal of the Royal Statistical Society. Series B (Methodological), 34(2), 187–202. 10.1111/j.2517-6161.1972.tb00899.x [DOI] [Google Scholar]
  21. Das, J., & Rajanikant, G. K. (2018). Post stroke depression: The sequelae of cerebral stroke. Neuroscience and Biobehavioral Reviews, 90, 104–114. 10.1016/j.neubiorev.2018.04.005 [DOI] [PubMed] [Google Scholar]
  22. Deyo, R. A., Cherkin, D. C., & Ciol, M. A. (1992). Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. Journal of Clinical Epidemiology, 45(6), 613–619. 10.1016/0895-4356(92)90133-8 [DOI] [PubMed] [Google Scholar]
  23. Dziewas, R., Beck, A. M., Clave, P., Hamdy, S., Heppner, H. J., Langmore, S. E., Leischker, A., Martino, R., Pluschinski, P., Roesler, A., Shaker, R., Warnecke, T., Sieber, C. C., Volkert, D., & Wirth, R. (2017). Recognizing the importance of dysphagia: Stumbling blocks and stepping stones in the twenty-first century. Dysphagia, 32(1), 78–82. 10.1007/s00455-016-9746-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Ekberg, O., Hamdy, S., Woisard, V., Wuttge-Hannig, A., & Ortega, P. (2002). Social and psychological burden of dysphagia: Its impact on diagnosis and treatment. Dysphagia, 17(2), 139–146. 10.1007/s00455-001-0113-5 [DOI] [PubMed] [Google Scholar]
  25. Fisher, L. D., & Lin, D. Y. (1999). Time-dependent covariates in the Cox proportional hazards regression model. Annual Review of Public Health, 20(1), 145–157. 10.1146/annurev.publhealth.20.1.145 [DOI] [PubMed] [Google Scholar]
  26. Goldmann, E., Roberts, E. T., Parikh, N. S., Lord, A. S., & Boden-Albala, B. (2016). Race/ethnic differences in post-stroke depression (PSD): Findings from the Stroke Warning Information and Faster Treatment (SWIFT) Study. Ethnicity & Disease, 26(1), 1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Hackett, M. L., & Pickles, K. (2014). Part I: Frequency of depression after stroke: An updated systematic review and meta-analysis of observational studies. International Journal of Stroke, 9(8), 1017–1025. 10.1111/ijs.12357 [DOI] [PubMed] [Google Scholar]
  28. Holland, G., Jayasekeran, V., Pendleton, N., Horan, M., Jones, M., & Hamdy, S. (2011). Prevalence and symptom profiling of oropharyngeal dysphagia in a community dwelling of an elderly population: A self-reporting questionnaire survey. Diseases of the Esophagus, 24(7), 476–480. 10.1111/j.1442-2050.2011.01182.x [DOI] [PubMed] [Google Scholar]
  29. Horn, J., Simpson, K. N., Simpson, A. N., Bonilha, L. F., & Bonilha, H. S. (2022). Incidence of poststroke depression in patients with poststroke dysphagia. American Journal of Speech-Language Pathology, 31(4), 1836–1844. 10.1044/2022_AJSLP-21-00346 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Hosmer, D. W., Lemeshow, S., & Sturdivant, R. X. (2013). Applied logistic regression (3rd ed.). Wiley. 10.1002/9781118548387 [DOI] [Google Scholar]
  31. Khedr, E. M., Abbass, M. A., Soliman, R. K., Zaki, A. F., & Gamea, A. (2021). Post-stroke dysphagia: Frequency, risk factors, and topographic representation: Hospital-based study. The Egyptian Journal of Neurology, Psychiatry and Neurosurgery, 57(1), Article 23. 10.1186/s41983-021-00281-9 [DOI] [Google Scholar]
  32. Kuehner, C. (2017). Why is depression more common among women than among men? Lancet Psychiatry, 4(2), 146–158. 10.1016/s2215-0366(16)30263-2 [DOI] [PubMed] [Google Scholar]
  33. Lazzarino, A. I., Hamer, M., Stamatakis, E., & Steptoe, A. (2013a). The combined association of psychological distress and socioeconomic status with all-cause mortality: A national cohort study. JAMA Internal Medicine, 173(1), 22–27. 10.1001/2013.jamainternmed.951 [DOI] [PubMed] [Google Scholar]
  34. Lazzarino, A. I., Hamer, M., Stamatakis, E., & Steptoe, A. (2013b). Low socioeconomic status and psychological distress as synergistic predictors of mortality from stroke and coronary heart disease. Psychosomatic Medicine, 75(3), 311–316. 10.1097/PSY.0b013e3182898e6d [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Leentjens, A. F., Aben, I., Lodder, J., & Verhey, F. R. (2006). General and disease-specific risk factors for depression after ischemic stroke: A two-step Cox regression analysis. International Psychogeriatrics, 18(4), 739–748. 10.1017/s1041610206003486 [DOI] [PubMed] [Google Scholar]
  36. Lin, D., Wei, L., & Ying, Z. (1993). Checking the Cox model with cumulative sums of martingale-based residuals. Biometrika, 80(3), 557–572. 10.1093/biomet/80.3.557 [DOI] [Google Scholar]
  37. Martino, R., Foley, N., Bhogal, S., Diamant, N., Speechley, M., & Teasell, R. (2005). Dysphagia after stroke: Incidence, diagnosis, and pulmonary complications. Stroke, 36(12), 2756–2763. 10.1161/01.STR.0000190056.76543.eb [DOI] [PubMed] [Google Scholar]
  38. Matthews, K. A., &, Gallo, L. C. (2011). Psychological perspectives on pathways linking socioeconomic status and physical health. Annual Review of Psychology, 62(1), 501–530. 10.1146/annurev.psych.031809.130711 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. McCarthy, M. J., Sucharew, H. J., Alwell, K., Moomaw, C. J., Woo, D., Flaherty, M. L., Khatri, P., Ferioli, S., Adeoye, O., Kleindorfer, D. O., & Kissela, B. M. (2016). Age, subjective stress, and depression after ischemic stroke. Journal of Behavioral Medicine, 39(1), 55–64. 10.1007/s10865-015-9663-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Mickey, R. M., & Greenland, S. (1989). The impact of confounder selection criteria on effect estimation. American Journal of Epidemiology, 129(1), 125–137. 10.1093/oxfordjournals.aje.a115101 [DOI] [PubMed] [Google Scholar]
  41. Moon, S., & Shin, J. (2006). Health care utilization among Medicare–Medicaid dual eligibles: A count data analysis. BMC Public Health, 6, Article 88. 10.1186/1471-2458-6-88 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Namasivayam-MacDonald, A. M., & Shune, S. E. (2018). The burden of dysphagia on family caregivers of the elderly: A systematic review. Geriatrics, 3(2), Article 30. 10.3390/geriatrics3020030 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group. (1995). Tissue plasminogen activator for acute ischemic stroke. The New England Journal of Medicine, 333(24), 1581–1588. 10.1056/NEJM199512143332401 [DOI] [PubMed] [Google Scholar]
  44. Ostir, G. V., Berges, I. M., Ottenbacher, A., & Ottenbacher, K. J. (2011). Patterns of change in depression after stroke. Journal of the American Geriatrics Society, 59(2), 314–320. 10.1111/j.1532-5415.2010.03266.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Paolucci, S., Iosa, M., Coiro, P., Venturiero, V., Savo, A., De Angelis, D., & Morone, G. (2019). Post-stroke depression increases disability more than 15% in ischemic stroke survivors: A case–control study. Frontiers in Neurology, 10, Article 926. 10.3389/fneur.2019.00926 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Paul, P., Pennell, M. L., & Lemeshow, S. (2013). Standardizing the power of the Hosmer–Lemeshow goodness of fit test in large data sets. Statistics in Medicine, 32(1), 67–80. 10.1002/sim.5525 [DOI] [PubMed] [Google Scholar]
  47. Pritchard, K. T., Hreha, K. P., & Hong, I. (2020). Dysphagia associated with risk of depressive symptoms among stroke survivors after discharge from a cluster of inpatient rehabilitation facilities. Swallowing Rehabilitation, 3(1), 33–44. 10.31115/sr.2020.3.1.33 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Quan, H., Li, B., Couris, C. M., Fushimi, K., Graham, P., Hider, P., Januel, J. M., & Sundararajan, V. (2011). Updating and validating the Charlson Comorbidity Index and score for risk adjustment in hospital discharge abstracts using data from 6 countries. American Journal of Epidemiology, 173(6), 676–682. 10.1093/aje/kwq433 [DOI] [PubMed] [Google Scholar]
  49. Quan, H., Sundararajan, V., Halfon, P., Fong, A., Burnand, B., Luthi, J. C., Saunders, L. D., Beck, C. A., Feasby, T. E., & Ghali, W. A. (2005). Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Medical Care, 43(11), 1130–1139. 10.1097/01.mlr.0000182534.19832.83 [DOI] [PubMed] [Google Scholar]
  50. Salk, R. H., Hyde, J. S., & Abramson, L. Y. (2017). Gender differences in depression in representative national samples: Meta-analyses of diagnoses and symptoms. Psychological Bulletin, 143(8), 783–822. 10.1037/bul0000102 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Saver, J. L., Fonarow, G. C., Smith, E. E., Reeves, M. J., Grau-Sepulveda, M. V., Pan, W., Olson, D. M., Hernandez, A. F., Peterson, E. D., & Schwamm, L. H. (2013). Time to treatment with intravenous tissue plasminogen activator and outcome from acute ischemic stroke. JAMA, 309(23), 2480–2488. 10.1001/jama.2013.6959 [DOI] [PubMed] [Google Scholar]
  52. Saxena, S. K., Ng, T. P., Yong, D., Fong, N. P., & Koh, G. (2008). Subthreshold depression and cognitive impairment but not demented in stroke patients during their rehabilitation. Acta Neurologica Scandinavica, 117(2), 133–140. 10.1111/j.1600-0404.2007.00922.x [DOI] [PubMed] [Google Scholar]
  53. Schober, P., & Vetter, T. R. (2018). Survival analysis and interpretation of time-to-event data: The tortoise and the hare. Anesthesia & Analgesia, 127(3), 792–798. 10.1213/ANE.0000000000003653 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Schreiber-Gregory, D. N. (2017). Multicollinearity: What is it, why should we care, and how can it be controlled? [White paper]. Proceedings of the SAS Global Forum 2017 Conference. https://support.sas.com/resources/papers/proceedings17/1404-2017.pdf
  55. Sheehan, T. J., Fifield, J., Reisine, S., & Tennen, H. (1995). The measurement structure of the Center for Epidemiologic Studies Depression Scale. Journal of Personality Assessment, 64(3), 507–521. 10.1207/s15327752jpa6403_9 [DOI] [PubMed] [Google Scholar]
  56. Sienkiewicz-Jarosz, H., Milewska, D., Bochyńska, A., Chełmniak, A., Dworek, N., Kasprzyk, K., Gałecka, K., Szczepańska-Szarej, A., Chwojnicki, K., Zyluk, B., Słowik, A., & Ryglewicz, D. (2010). Predictors of depressive symptoms in patients with stroke—A three-month follow-up. Neurologia i Neurochirurgia Polska, 44(1), 13–20. 10.1016/s0028-3843(14)60402-3 [DOI] [PubMed] [Google Scholar]
  57. Simpson, A. N., Wilmskoetter, J., Hong, I., Li, C. Y., Jauch, E. C., Bonilha, H. S., Anderson, K., Harvey, J., & Simpson, K. N. (2018). Stroke Administrative Severity Index: Using administrative data for 30-day poststroke outcomes prediction. Journal of Comparative Effectiveness Research, 7(4), 293–304. 10.2217/cer-2017-0058 [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Stewart, J. T. (2004). Why don't physicians consider depression in the elderly? Age-related bias, atypical symptoms, and ineffective screening approaches may be at play. Postgraduate Medicine, 115(6), 57–59. 10.3810/pgm.2004.06.1539 [DOI] [PubMed] [Google Scholar]
  59. Stoltzfus, J. C. (2011). Logistic regression: A brief primer. Academic Emergency Medicine, 18(10), 1099–1104. 10.1111/j.1553-2712.2011.01185.x [DOI] [PubMed] [Google Scholar]
  60. Towfighi, A., Ovbiagele, B., El Husseini, N., Hackett, M. L., Jorge, R. E., Kissela, B. M., Mitchell, P. H., Skolarus, L. E., Whooley, M. A., & Williams, L. S. (2017). Poststroke depression: A scientific statement for healthcare professionals from the American Heart Association/American Stroke Association. Stroke, 48(2), e30–e43. 10.1161/STR.0000000000000113 [DOI] [PubMed] [Google Scholar]
  61. UCLA: Statistical Consulting Group. (n.d.). Introduction to survival analysis in SAS. Retrieved July 10, 2023, from https://stats.idre.ucla.edu/sas/seminars/sas-survival/
  62. Verdonschot, R. J., Baijens, L. W. J., Serroyen, J. L., Leue, C., & Kremer, B. (2013). Symptoms of anxiety and depression assessed with the Hospital Anxiety and Depression Scale in patients with oropharyngeal dysphagia. Journal of Psychosomatic Research, 75(5), 451–455. 10.1016/j.jpsychores.2013.08.021 [DOI] [PubMed] [Google Scholar]
  63. Verdonschot, R. J., Baijens, L. W. J., Vanbelle, S., van de Kolk, I., Kremer, B., & Leue, C. (2017). Affective symptoms in patients with oropharyngeal dysphagia: A systematic review. Journal of Psychosomatic Research, 97, 102–110. 10.1016/j.jpsychores.2017.04.006 [DOI] [PubMed] [Google Scholar]
  64. Zhang, S., Xu, M., Liu, Z. J., Feng, J., & Ma, Y. (2020). Neuropsychiatric issues after stroke: Clinical significance and therapeutic implications. World Journal of Psychiatry, 10(6), 125–138. 10.5498/wjp.v10.i6.125 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The data sets generated and analyzed during the current study are not publicly available due to license restrictions.

Articles from American Journal of Speech-Language Pathology are provided here courtesy of American Speech-Language-Hearing Association

RESOURCES