Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 Dec 1.
Published in final edited form as: J Technol Behav Sci. 2017 Nov 28;2(3-4):129–139. doi: 10.1007/s41347-017-0025-3

Using Telemedicine to Identify Depressive Symptomatology Rating Scale in a Home Parenteral Nutrition Population

Natasia Adams 1, Nancy Hamilton 2, Eve-Lynn Nelson 3, Carol E Smith 4
PMCID: PMC5978432  NIHMSID: NIHMS923451  PMID: 29862312

Introduction

Nutrition support therapies are required for multiple underlying pathologies that cause malnutrition and nutrient imbalance (Kirby 2012). Home Parenteral Nutrition (HPN), a nutrition infusion therapy, addresses the nutrient imbalance specifically caused by a malfunctioning gastrointestinal tract and increases life expectancy outside of a medical inpatient care setting and within a home environment (Baxter et al. 2006; Huisman-de Waal et al. 2007; Persoon et al. 2005). Limited epidemiological data exist regarding HPN use within the United States. However, previous data collected from 1989 to 1992 showed that 120 per 100,000 individuals utilized HPN (DiBaise and Scolapio 2007). In a more recent survey in 2010, the National Center for Health Statistics approximated that 33,000 individuals, across the lifespan, received HPN (Winkler et al. 2015). With the cost of HPN calculated to be around $652 million, HPN causes a financial burden on the families and the health care system (Winkler et al. 2015). The financial burden is compounded further when other complications associated with HPN arise. For instance, extant literature has shown that HPN patients’ experience not only physical complaints (e.g., diarrhea) but also psychosocial concerns related to their treatment. While majority of the physical concerns can be resolved in the home environment, HPN patients’ psychosocial concerns, specifically depression, are left unaddressed (Baxter et al. 2006; Huisman-de Waal et al. 2006; Persoon et al. 2005). Depression as a psychosocial concern accrues an additional cost (around $80 billion) and reduces quality of life (Howard 2006). Thus, to address depression within the HPN population, it is imperative that healthcare professionals are able to detect depressive symptoms as they appear, so they may offset healthcare cost, reduce psychosocial concern, and improve quality of life among HPN users.

Detecting and addressing both physical and psychosocial concerns can be difficult with HPN users, particularly those receiving treatment outside of a hospital setting. The advent of telemedicine as a potential mode of treatment has allowed healthcare professionals to treat patients in the home environment. Telemedicine increases HPN users’ access to healthcare professionals (e.g., mental health specialists) and vital health care resources. Thus, telemedicine reduces barriers to medical care. However, questions remain about whether the absence of face-to-face contact may limit the ability of healthcare professionals to detect psychological problems, like depression, particularly in chronically ill population. Therefore, in the current study, we investigate whether it is feasible to assess depressive symptomatology during a telemedicine encounter with patients in the long-term HPN population.

As stated above, quality of life in the HPN population is considerably reduced compared to the disease-free population (Huisman-de Waal et al. 2007; Persoon et al. 2005). Patients receiving HPN frequently report exhaustion, “loss of energy and optimism, fear”, apprehension due to the future, and concern about the underlying disease (Huisman-de Waal et al. 2007; Winkler and Smith 2014). HPN patients also reported a loss of independence and loss of engagement in previous activities (Baxter et al. 2006; Huisman-de Waal et al. 2006, 2007; Persoon et al. 2005). Although seen infrequently, female HPN users have reported concerns of body image due to the ports and tunneled catheters implanted in the body (Baxter et al. 2006; Huisman-de Waal et al. 2007; Winkler and Smith 2014). These concerns may reflect the presence of or contribute to the high prevalence of depression reported by patients with HPN (Baxter et al. 2006; Huisman-de Waal et al. 2006, 2007; Persoon et al. 2005; Winkler and Smith 2014; Winkler 2005; Winkler and Guenter 2014). Another study conducted by Howard (2006) discovered a reduction in quality of life in those who rapidly transition from good health to using HPN and in those who are just beginning HPN treatment. According to the author, factors that lowered quality of life were the following: “loss of employment, loss of income, and decreased social interaction (Howard 2006).” These factors found in the study certainly contribute to negative mood states leading to the increased diagnosis of depression disorder within the HPN population. Consequently, HPN users are at a higher risk to experience reduced quality of life and a depressive episode due to HPN therapy than their healthy counterparts.

Since long-term HPN patients face an increased risk of depression, healthcare providers should aim to increase monitoring and identifying the presence of depressive symptomology within this population. Unfortunately, long-term HPN patients are commonly restricted to the home environment and experiences reduced access to mental healthcare providers. Alternative avenues outside of traditional modes of treatment need to be used to assess for depressive symptomatology. An available alternative avenue is telemedicine. Telemedicine allows long-term HPN patients to have increased access to healthcare providers and the essential services they provide. Moreover, investigators have demonstrated that telemedicine is comparable in its effectiveness to treat depression to traditional mode of therapy (i.e., face-to-face therapy) (De Las Cuevas et al. 2006; Khatri et al. 2014;O’Reilly et al. 2007). With long-term HPN patients designated by the federal government as being a technology dependent because of their reliance on technology (i.e., infusion pumps) for survival, this population has been primed to accept the introduction of telemedicine, which should provide treatment support. Indeed, researchers investigate the use of telemedicine within the HPN population have shown HPN patients to be amendable to utilizing videoconferencing as another avenue for treatment (Saqui et al. 2007; Smith et al. 2015). This finding is supported in a recent study (Smith et al. 2015), where researchers found HPN patients as well as healthcare professionals willing to communicate utilizing videoconferencing through the use of mobile health technology (Smith et al. 2015). Using telemedicine to identify and monitor depressive symptoms within this population is a feasible solution for both healthcare professionals and HPN users.

To accurately access the feasibility of utilizing telemedicine to identify depressive symptomatology compared to traditional modes (face- to- face), a framework that assesses the effectiveness of telemedicine interventions was used to guide the current study. The Model for Assessment of Telemedicine (MAST) is a popular framework used in Europe to evaluate the application of telemedicine interventions (Kidholm et al. 2012). The MAST framework has three major components: (1) preceding consideration, (2) multidisciplinary assessment, and (3) assessment of transferability. The preceding consideration component comprises of identifying the purpose in which the technology shall be used and if the technology is sufficiently developed to support the intervention. The multidisciplinary component includes the following domains (1) health problem and characteristics of the application, (2) safety, (3) clinical effectiveness, (4) patient perspectives, (5) economic aspects, (6) organizational aspects, and (7) social-cultural, ethical, and legal aspects. The last component transferability refers to extrapolating the intervention to other settings and populations. The current study meets the preceding consideration criteria as we have demonstrated a clear purpose for utilizing developed technological resources on a national level. The focus of the current study explores in depth the multidisciplinary component, specifically the clinical effectiveness of utilizing of telemedicine to identify symptoms of depression.

Extant research has shown telemedicine being effective in treating depression in different medical populations (e.g., pain, diabetes, and disability); however, limited evidence exists in identifying depressive symptoms using telemedicine versus traditional diagnostic tools, such as structured interviews and self-report measures (Baron et al. 2016; Chavooshi et al. 2016; Choi, et al. 2014; Mochari-Greenberger 2016; Salisbury, et al. 2016). To assess whether telemedicine can identify depressive symptoms, the current study has modified an existing validated scale, the Raskin Three- Area Severity of Depression rating scale. Currently, the modified scale is the best option to detect depressive symptoms through telemedicine since it rates the observation of depressive symptoms and allows for easier detection of depressive symptoms within an unstructured interview. Hence, using the modified scale allows the current study to explore not only whether depressive symptoms can be in identified through telemedicine but also add to the existing literature regarding the efficacy of telemedicine versus traditional diagnostic tools.

As such, the aims of the study were to assess interrater reliability of the modified Raskin scale in rating depressive symptoms; to establish criterion validity of the scale; and to identify items on the modified Raskin scale that correlated either strongly or poorly with the PHQ-9 scores within a long-term HPN population. We hypothesized that raters would demonstrate good interrater reliability and the modified Raskin scale would display excellent criterion validity in assessing depressive symptoms within the long-term HPN population. We also hypothesized that secondary symptoms from the modified Raskin scale would correlate poorly and behavior symptoms and verbal report from the modified Raskin scale would correlate strongly with the PHQ-9 and allows a better understanding of which symptoms would be indicative of depression over a video-supported platform.

Methods

The dataset presented in the current study was procured from a clinical trial study proposed by Carol Smith, Ph.D. and funded by the National Institutes of Health (NIH) and registered on the USA ClinicalTrials.gov as Registration # NCT0190028. The NIH clinical trial enrolled 126 participants to examine the feasibility of telemedicine (mobile technology) to provide health care services to long-term HPN users and caregivers as well as assist this population with self-care and coordinating that care with other healthcare professionals. The Institutional Review Board at the University of Kansas Medical Center approved the original study and confidentiality has been kept with all patient information (i.e., videos) kept on a secure drive.

Participants

Researchers recruited participants through the University of Kansas Medical Center, associated rural health centers, and the Oley Foundation. Participants were eligible to participate if they were HPN user and had a caregiver (non HPN- user) who agreed participate in the study, fluent in English, and able to provide informed consent. Exclusion criteria included any participants that had a disability or disorder that prevented them from using a mobile tablet. In the current study, participants were excluded if they were below the age of 18 (n=2), did not participate in the facilitated group discussion (n=2), were not a HPN user (n=38) or in the control group (n=59).

Questionnaires

Demographic Questionnaire

Participants completed a demographic questionnaire that collected information about income, insurance, and participants’ use of HPN (e.g., reasons for HPN and duration of HPN).

Patient Health Questionnaire-9 (PHQ-9)

PHQ-9 (Kroenke et al. 2001) is a nine- item questionnaire that asks respondents to respond to items such as I have “little interest or pleasure in doing things” on a four- point, Likert scale. The PHQ-9 was created to monitor and/measure the severity of depressive symptomatology and has shown excellent internal reliability (alpha = 0.89) and test-retest reliability. The PHQ-9 has also shown excellent discrimination between individuals with depression and individuals without depression. Kroenke, et al. (2001) demonstrated that the PHQ-9 has good criterion validity with regard to its ability to identify and predict depressive symptoms. Kroenke, et al. also demonstrated good construct validity with the PHQ-9 strong relationship between “functional status, disability days, and symptom-related difficulty” (2001).

Raskin Three- Area Severity of Depression Rating Scale (Raskin scale)

Before selecting the Raskin scale, the author reviewed the literature and consulted with depression experts on depressive measures (e.g., Beck Depression Inventory-II, Montgomery-Asberg Depression Rating) that would allow for an observational assessment of depressive symptoms through an unstructured interview. The majority of the traditional depressive measures for adults relied on either self-report or a structured/semi-structured interview conducted by the clinician to identify depressive symptomatology. However, the current study required a measure that could be used to rate symptoms of depression as they presented themselves in an unstructured group setting, rather than in a structured interview. The Raskin scale (Raskin et al. 1970) was designed as an observer rater scale to measure changes in depressive symptoms of depressed individuals undergoing pharmacological treatment. The Raskin scale examines depression symptoms in three major areas—verbal report (e.g., feels blue, reports of crying spells), behavior (e.g., looks sad, lacking energy), and secondary symptoms (e.g., insomnia, change in appetite) with examples anchoring in each area. Raters rate the degree of severity from one (not at all) to five (very much), for each area. Thus, scores can range from three to 15 points to demonstrate the severity of the depression. Psychometric properties of the measure are limited; however, studies have shown the Raskin to have acceptable interrater reliability (.88 for the entire scale), good sensitivity but poor specificity (Fish 2011; Nezu, et al. 2000). The Raskin scale was selected because it differs from self-report and interview measures in that multiple sources of information (i.e., interview, patient self-report, or collateral information provided by others) are used to inform the assessment of depression symptom. This may make the Raskin the best option for an observer rating scale when interview data are not available.

The Raskin scale (see Appendix A) was modified to identify the presence of depression rather than the severity of a depressive mood. When modifying the Raskin scale, the author added depressive symptoms specified by the Diagnostic Statistical Manual- 5 (DSM-5) and from the existing literature (Cummins et al. 2015; Foley and Gentile 2010). The author added the following symptoms in the Verbal Report domain: limited speech production and complaints of loss energy or fatigue; in the Behavioral Symptom domain: poor eye contact and flat facial expression; and Secondary Symptom domain: self-focus, previously reported depression and optional symptoms (i.e., sleep disturbances). An additional modification was made to the rating scale. Instead of rating each domain on a Likert scale, raters were asked to code the presence of the symptom under each domain by marking the presence of a symptom, a yes (1) or no (0), and tabulating the number of times the symptom occurred. Thus, the modified Raskin score could have ranged from 0–294. In practice, however, the items on the modified Raskin scale ranged from 0–4.

Procedure

As part of the original study, participants were sent a wireless touch-screen mobile Apple iPad Mini™ tablet (iPad). The iPad included a fourth generation (4G) unlimited data plan and a five mega-pixel camera that provided clear visual properties for participants. Polycom RealPresence encrypted software was installed onto the iPads to allow for secure videoconferencing between health professionals and participants. Participants completed a PHQ-9 questionnaire and engaged in a facilitated group discussion at baseline (T1). Participants then completed another PHQ-9 questionnaire a month after the facilitated group discussion (T2).

To assess whether depressive symptoms could be identified through videoconferencing, three raters reviewed 28 facilitated group discussions that were videotaped by an iPad. Raters were graduate students from the University of Kansas Clinical Psychology program with clinical experience in identifying depression. Raters underwent training in using the modified Raskin Scale and were blinded to both PHQ-9 scores at T1 and T2. With the modified Raskin scale, raters recorded the presence of a specific symptom (e.g. poor eye contact) with a yes or no and tabulated the number the times that specific depressive symptom occurred. If raters reported difficulty with identifying or tabulating the presence of a depressive symptom, the other two raters were consulted. Videotapes were reviewed and discussed until all raters arrived at a consensus about the presence/absence of a depressive symptom.

Data Analysis

Statistical analyses were conducted using the Statistical Package for Social Sciences -22 (SPSS-22) software packages. To examine interrater reliability, interclass correlation coefficients (ICC) were calculated (Koo and Li 2016; Shrout and Fleiss 1979). Criterion validity was assessed by comparing the modified Raskin scale to the PHQ-9 at T1 and T2. An overall sum was created for the modified Raskin scale. An OLS was conducted to determine whether the modified Raskin scale correlated with patient reported depression symptoms on the PHQ-9 score at T1 and T2. Zero order correlations were examined to evaluate the relationship between subscales of the Raskin—Verbal Report, Behavioral Symptoms, and Secondary Symptoms and the total PHQ-9 score (T1 and T2).

Results

Characteristics of the sample

Twenty-five patients between the ages of 19–71 (M=41.76, SD= 14.36) were included in the analyses at T1. Participants were predominantly female (88%), Caucasian (96%), and were on HPN for at least two years (44%). See Table 1 for further description of the baseline demographics of the sample. Four patients out of the 25 patients from T1 were excluded from analyses at T2 for the following reasons: three patients did not complete the PHQ-9 questionnaire at T2, and one participant died. We found no differences between those who completed the study and those four who dropped out of the study.

Table 1.

Characteristics of Long-Term HPN Users

Characteristics n(%) M (SD)
Sex
 Males 3 (12%)
 Females 22 (88%)
Ethnicity
 Hispanic 2 (8%)
 Non-Hispanic 23 (92%)
Race
 White 24 (96%)
 Black 1 (4%)
Age 41.76 (14.36)
HPN Years 2.76 (.926)
 1 1 (4%)
 2 11 (44%)
 3 6 (24%)
 4 7 (28%)
Education
 Some high school/Currently in high school 1 (4%)
 Some college/Currently in college 6 (24%)
 Completed college or more 18 (72%)
Marital Status
 Married 9 (36%)
 Divorced 5 (20%)
 Separated 1 (4%)
 Never Married 10 (40%)
Household (#of people in home including yourself)
3.56 (2.256)
 1 person 3 (12%)
 2 people 8 (32%)
 3 people 8 (32%)
 4 people 4(16%)
 5 people 2 (8)
Modified Raskin Scale 3.28(3.458)
PHQ-9 at T1 25 7.88 (5.464)
PHQ-9 at T2 21 7.76 (5.638)
Open in a new tab

Note. HPN Years= Home Parenteral Nutrition; PHQ-9= Patient Health Questionnaire -9

To assess the interrater reliability of the modified Raskin scale, three clinical psychology graduate students rated three long-term HPN users on the modified Raskin scale after training. Interrater reliability was calculated using ICC. We used a two-way random effects model to examine the average measurement and absolute agreement between the raters (Koo and Li 2016). The interrater reliability was found to be good to excellent amongst raters, ICC (2, 3) = 0.96 with a 95% confidence interval=0.83–0.999, F (2, 8) = 3.03, p<.05.

A follow-up analysis examined whether the difference between T1 and T2 was due to the loss of the four patients from T1. We removed the four patients lost at T2 and re-ran the OLS regression equation at T1 to determine whether that changed the relationship between the modified Raskin scale and PHQ-9 at T1. The results were essentially unchanged. The modified Raskin scale explained 3.5 % of the variance [R2=.035, F (1, 19) =.681, p=.419] in the PHQ-9 at T1. Thus, the difference in the relationship between the Raskin scale and PHQ-9 at T1 and T2 was not an artifact of patient attrition.

An OLS regression analysis was used to determine whether the total of the modified Raskin scale predicted self-reported depression symptoms as identified by the PHQ-9 at T1 (Table 2). The results indicated that the modified Raskin scale explained .8% of the variance [R2=.008, F (1, 23) =1.88, p =.669] in the PHQ-9 at T1. In other words, there was no relationship between the modified Raskin scale and PHQ-9. Thus, the modified Raskin scale did not predict the presence of depression as measured by the PHQ-9. A second OLS analysis was performed to determine whether the modified Raskin scale predicted the presence of depression as identified by the PHQ-9 scores at T2. The results indicated that the modified Raskin scale explained 46.5% of the variance [R2=.465, F (1, 19) = 16.486, p< .01]. It was found that the modified Raskin scale significantly predicted self-reported depression symptoms as identified by the PHQ-9 (β=.682, p<.01).

Table 2.

Summary of Ordinary Least Square Regression Analyses for the Modified Raskin Scale Predicting PHQ-9 at T1 and T2

Variable PHQ-9
T1 (n=25) T2 (n=21)
β t P 95% CI β t p 95% CI
Constant 4.791 .000 [4.21, 10.61] 3.336 .003 [1.58, 6.89]**
Modified Raskin Scale .090 .434 .669 [−0.54, 0.82] .682** 4.060 .001 [0.51, 0.58]**
Open in a new tab
*

p<.05,

**

p<.01

Finally, in a second follow-up analysis we used an OLS regression analysis to determine whether the PHQ-9 at T1 or the modified Raskin scale at T1 predicted depressive symptoms at T2 after controlling for PHQ-9 at T1 (Table 3). The results indicated that PHQ-9 at T1 and the modified Raskin scale accounted for 67.9% of the variance [R2=.679, F (2, 18) = 18.997, p<.01)]. Both the PHQ-9 at T1 (β=.471, p<.01) and the modified Raskin scale (β=.594, p<.01) significantly predicted depressive symptoms as identified by the PHQ-9 at T2.

Table 3.

Summary of Ordinary Least Square Regression Analyses for the Modified Raskin Scale and PHQ-9 at T1 Predicting PHQ-9 at T2

PHQ-9 at T2 (n=21)
Variables β t P 95% CI
Constant .774 .449 [−1.810, 3.923]
PHQ-9 at T1 .471 3.46 .003 [0.193, 0.790]**
Modified Raskin Scale .594 4.37 .000 [0.472, 1.348]**
Open in a new tab

Note. PHQ-9 at T1: R2Δ= .338; Modified Raskin Scale: R2Δ=3.41;

*

p<.05,

**

p<.01

Pearson correlations were calculated between the domains—Verbal Report (VR), Behavioral Symptoms (BS), and Secondary Symptoms (SS)—of the modified Raskin scale and the PHQ-9 scores at T1 and T2 (Table 4). Nonsignificant Pearson correlations were found between all the domains of the Raskin and PHQ-9 at T1 (p=ns). However, significant Pearson correlations were found between the PHQ-9 at T2 and two of the Raskin subscales at T2 VR (r=.463) and SS (r=.659) and PHQ-9 at T2.

Table 4.

Correlation of Domains of the Modified Raskin Scale, Modified Raskin Scale Total, and PHQ-9 at T1 and T2

Variables M(SD) 1 2 3 4 5 6
1. PHQ-9 at T1 7.88(5.46) –
2. PHQ-9 at T2 7.76(5.64) .58** –
3. Verbal Report .84(1.60) .01 .46* –
4. Behavioral Symptoms .56(1.12) .33 .38 −.09 –
5. Secondary Symptoms 1.88(1.90) −.04 .66** .71** .13 –
6. Modified Raskin Scale 3.28(3.46) .09 .68** .82* .36 .92** –
Open in a new tab
*

p<.05,

**

p<.01

Follow-up analyses were conducted to identify specific Raskin items that correlated with the self-reported depression. Specifically, we examined zero-order correlations between the PHQ-9 at T1 and T2 and individual items from the modified Raskin scale (Table 5). Significant relationships exist between PHQ-9 at T1and “appears restless” (BS) (r=.405) and “previously reported depression” (SS) (r=−.509). Significant correlations exist between the PHQ-9 at T2 and “complains of loss energy or fatigue” (VR) (r=.474), the “complaints of insomnia” (SS) (r=.457), “self-focus” (SS) (r=.514), and optional symptoms (i.e., sleep disturbances) (SS) (r=.531).

Table 5.

Correlation of Individual Items on the Modified Raskin Scale with the PHQ-9 at T1 and T2

Variables M(SD) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
1. Limited speech production 0(0) . . . . . . . . . . . . . . . . . . . . . . .
2. Participant reports feelings of Hopelessness/Worthlessness/Guilt .12 (.33) . 1 . .67** . . .70** −.08 . −.13 . . . −.10 −.01 −.08 . .56** .52** .39 .36 −.02 .31
3. Participant complains of losing interest or pleasure in activities 0(0) . . . . . . . . . . . . . . . . . . . . . . .
4. Reports Depressed Mood .16(.47) . .67** . . . . . . . . . . . −.01 −.09 −.07 . .81** .64** .57** .46* −.22 .38
5. Wishes to be Dead 0(0) . . . . . . . . . . . . . . . . . . . . . . .
6. Reports Crying Spells . . . . . . . . . . . . . . . . . . . . . . . .
7. Complains of loss energy or fatigue .56(.96) . .70** .71** . . 1 −.12 . −.20 . . . .23 −.16 .31 . .75** .84** .28 .47* .12 .47*
8. Looks Sad .08(.40) . −.08 . −.07 . . −.12 1 . −.07 . . . −.06 −.05 −.04 . −.04 −.07 −.18 −.18 .04 0
9. Crying 0(0) . . . . . . . . . . . . . . . . . . . . . . .
10. Poor Eye Contact .24(.72) . −.13 −.12 −.20 −.07 1 −.09 .01 −.07 −.07 −.12 −.13 −.11 −.00 .13
11. Speaks in low voice (Depressed Mood) 0(0) . . . . . . . . . . . . . . . . . . . . . . .
12. Movements are slow 0(0) . . . . . . . . . . . . . . . . . . . . . . .
13. Flat Facial Expression 0(0) . . . . . . . . . . . . . . . . . . . . . . .
14. Appears Restless .24(.88) . −.10 . −.10 . . .23 −.06 . −.09 . . . 1 .38 .89** . −.06 .56** −.24 −.04 .41* .38
15. Complaints of Insomnia .16(.62) . −.10 . −.09 . . −.16 −.05 . .10 . . . .38 1 −.05 . −.05 −.09 −.23 −.16 .19 .46*
16. Complaints of Hypersomnia .08(.40) . −.08 . −.07 . . .312 −.04 . −.07 . . . .89** −.05 1 . −.04 .65** −.18 .05 .39 .38
17. Reports recent change in appetite 0(0) . . . . . . . . . . . . . . . . . . . . . . .
18. Participant reports diminished ability to think or indecisiveness .04(.20) . .55** . .81** . . .75** −.04 . −.07 . . . −.06 −.05 −.04 . l .65** .41* .27 −.15 .34
19. Self-Focus .20(.58) . .52** . .64** . . .84** −.07 . −.12 . . . .56** −.09 .65** . .65** 1 .20 .32 .14 .51*
20. Previously reported depression .60(.71) . .39 . .57** . . .28 −.18 . −.13 . . . −.24 −.23 −.18 . .41* .20 1 .58** .51** .20
21. Optional symptoms (e.g., sleep disturbance) .80(.91) . .36 . .46* . . .47* −.18 . −.11 . . . −.04 −.16 .05 . .27 .32 .581** 1 −.04 .53*
22. PHQ-9 at T1 7.89(5.46) . −.02 . −.22 . . .12 .04 . −.003 . . . .41* .19 .39 . −.15 .14 −.51** −.04 l .58**
23. PHQ-9 at T2 7.76(5.64) . .31 . .38 . . .47* . . .129 . . . .38 .46* .38 . .34 .51* .20 .53* .58** 1
Open in a new tab

Notes. N was 25 for every variable except for PHQ-9 at T2, which N was 21 due to missing data. .=no value.

*

p<.05,

**

p<01

Discussion

HPN is a treatment that allows patients to receive competent care in the comfort of their own home, improving quality of life. However, patients on HPN lack the daily in person interaction with healthcare professionals, which may cause delay in identifying symptoms of depression by healthcare professionals. Thus, the purpose of the current study was to determine whether symptoms of depression could be identified over a video-supported platform utilizing an observer rating scale (i.e., the modified Raskin scale) as a way to avoid the delay caused by the lack of face-to-face interactions between patients and healthcare professionals. The results of this study suggest strong interrater reliability and mixed preliminary evidence regarding the criterion validity of the modified Raskin scale in identifying depressive symptoms over a video-supported platform compared to the PHQ-9.

The first AIM of this study was to assess the reliability of the Raskin scale. The interrater reliability as assessed by the ICC indicated that the raters were good to excellent in identifying depressive symptoms utilizing the modified Raskin scale, which supported our initial hypothesis. Documenting inter-rater reliability for an observer rating scale is of critical importance because the scale relies on the subjective interpretation and judgement of multiple raters. As such, multiple raters introduce measurement error, which negatively affects how well an observer rating scale assesses the intended construct (e.g., depression) (Hallgren 2012). We can reasonably assume that our use of multiple raters did not add excessively to the measurement error and compromise our ability to identify depressive symptoms using the modified Raskin scale.

The second and third AIMs of this study were to establish evidence of criterion validity by comparing the Raskin rating scale to an established self-report measure of depression symptoms. We predicted that the modified Raskin scale would be comparable to the self-report measure, PHQ-9, in detecting depressive symptoms. Unfortunately, there was little relationship between the total Raskin scale and the PHQ-9 at T1, indicating that these two measures of depression may not be measuring the same thing. Subdomain analysis proved similarly disappointing results. We had predicted that the two most objective domains of the Raskin that were based on observations of behaviors and verbal reports would have a stronger relationship to the PHQ-9 than the more subjective, secondary symptom domain. It is noteworthy, however, that although the relationship between behavioral symptoms and the PHQ-9 was not statistically significant, the effect size was much larger than the effect sizes of the relationship between the PHQ- 9 and verbal reports or secondary symptoms (Refer to Table 4).

Interestingly, although there was no relationship between the Raskin subscales and the PHQ-9 at Time 1, we found that the modified Raskin scale was able to predict depressive symptoms when compared to the PHQ-9 at T2. Hence, our findings suggest that while raters demonstrated excellent interrater reliability the modified Raskin scale demonstrated evidence of poor concurrent validity and strong predictive validity. To ensure that this was not an artifact because of the loss of patients in T2, we re-analyzed the data at T1 excluding the four patients lost at T2. We found the same non-relationship between the modified Raskin and the PHQ-9 at T1 that was seen in our earlier analysis, meaning that the loss of patients was not an artifact that caused our significant finding at T2.

There are several explanations for our mixed findings. One possible explanation is that the psychometric properties of the original Raskin scale are still evident in the modified Raskin scale. For instance, although the original Raskin scale had good sensitivity, the scale demonstrated poor specificity. Thus, our ability to differentiate depressive symptoms from common symptoms of HPN, such as inattentiveness due to pain or fatigue, may be compromised. Indeed, raters may mistakenly code illness behaviors as indicative of depression, which may be affecting the measure’s concurrent validity.

An alternative explanation could be that the PHQ-9 overestimates the presence of depression symptoms. The current sample was comprised of patients with critical illnesses that necessitate the use of HPN. Long-term HPN users experience physical symptoms (e.g., fatigue and weight loss) as result of their treatment that are not due to psychological cause. For example, the PHQ-9 has three questions that ask specifically about fatigue, appetite, and sleep. However, long-term HPN users are unlikely to be able to distinguish the presence of a symptoms caused by HPN as opposed to a psychological origin. Because we did not have a gold-standard diagnostic interview, it is not possible to determine which measure over or underestimated the presence or severity of depression. Additionally, we must consider the impact of the group dynamics and its influence on rating the PHQ-9 at T1 and T2. HPN patients might have felt more comfortable in disclosing depressive symptoms on the PHQ-9 at T2 as they became comfortable with the study’s team.

Consistent with AIM 3, we examined the correlation between the Raskin subscales and the PHQ-9 total score. Given that we were comparing observations to self-reports, we hypothesized that the more objective Raskin scale domains, VR and BS, would have stronger correlations with the PHQ-9 than the more subjective secondary symptom subscale. Consistent with the overall score, we saw the same pattern at T1—weak non- significant concurrent relationships between the modified Raskin subscales and the PHQ-9. However, we saw significant correlations between the modified Raskin scale domains to the PHQ-9 at T2. Surprisingly, the pattern of relationships was not as expected. The strongest relationship was between SS and the PHQ-9 at T2. VR was also a significant predictor of the PHQ-9 at T2. However, the BS domain had a non-significant relationship with the PHQ-9 at T2. We expected the behavioral domain to be easier to rate than the other domains. However, raters reported difficulty in rating behavioral symptoms because of the nature of the video-supported platform. Meaning, raters noted technical difficulties (i.e., poor video quality, constant moving of mobile tablet) that hindered their ability to successfully rate behavioral symptoms. In contrast, VR and SS relied on participants reporting symptoms, which were not influenced by video quality. Participants mostly reported sleep disturbances, fatigue related to their use of HPN. They also discussed weight loss and gain in reference to their use of HPN. Participants did talk about depressive symptoms. However, they usually reported such symptoms that had occurred in the past. Thus, VR and SS were easier to rate than BS, which may explain why VR and SS had a positive, significant relationship, but BS was not related to PHQ-9 at T2.

Overall, the modified Raskin scale demonstrated a poor ability to identify depressive symptoms compared to the PHQ-9 when administered simultaneously at T1. Hence, the PHQ-9 an inexpensive self-report measure seems to be a more appropriate tool than an observer rating scale—a disappointing finding considering that self-report measure may not be feasible when interactions occur over a video-supported platform. However, we found that the modified Raskin scale has utility in predicting depressive symptoms at T2, specifically, the VR and SS domains of the modified Raskin scale. Given these findings, it seems appropriate to explore whether the modified Raskin scale could be refined to increase its utility in identifying people at risk for increased symptoms of depression or onset of an episode of depression.

Subsequent analysis identified a few significant relationships between individual items on the modified Raskin scale and the PHQ-9 at T1 and T2 (Refer to Table 5). Although a weak relationship existed between the modified Raskin scale and the PHQ-9 at T1, it is interesting to note that, “appears restless” and “previously reported depressed,” had a strong, positive relationship with the PHQ-9 at T1. T1 observations: “complains of loss energy or fatigue”, “complaints of insomnia”, and behavioral observations of “self-focus”, and “optional symptoms”—demonstrated a strong relationship with the PHQ-9 at T2.

In examining individual items from the modified Raskin scale, we found it noteworthy that “self-focus” demonstrated a strong relationship with depressive symptoms as identified by the PHQ-9 at T2. Introduced by social psychologists in the seventies, self-focused attention has been documented in the literature to be closely related to depression (Ingram 1990; Mor and Winquist 2002; Pyszcynski et al. 1989). Indeed, researchers have argued that depressed individuals engage in self-focused attention as a form of rumination, which is characteristic of depressed individuals. Self-focused attention as a form of rumination appeared to be supported by our current findings as self-focus predicted that HPN patients, who focused more on themselves at T1, would exhibit an increase in depressive symptomatology in T2.

Our results suggest that the modified Raskin scale was sensitive to self-reports of vegetative symptoms (e.g., insomnia) and behavioral-related depressive symptoms (e.g., talking only about themselves) reported by patients. The data also suggest that these items from the modified Raskin scale were strong predictors of depression at T2. As such, the modified Raksin scale could be reduced to these items to lessen the observation burden. Although it may be beneficial for patients to continue self-reports of depressive symptoms at baseline, using the modified Raskin scale may provide its greatest utility in predicting future depressive symptomology. Meaning, healthcare professionals may identify individual patients, who may not directly report these behaviors (e.g., vegetative, behavioral-depressive symptoms) at baseline, but would warrant further monitoring and support in treating depression symptoms in the future.

Surprisingly, the modified Raskin scale was not only a predictor of future depressive symptoms as indicated by the PHQ-9 score at T2 but a stronger predictor than the PHQ-9 score at T1 (Refer to Table 3). This is an interesting finding considering that the PHQ-9, an empirically validated measure of depression, was expected to be a stronger predictor of itself (PHQ-9 score at T2) than the modified Raskin scale. We posit that the modified Raskin scale as an observer rating scale allows professionals to identify vegetative, behavioral-depressive symptoms that are more indicative of a future experience of depression. Although PHQ-9 consists of questions regarding vegetative, behavioral depressive symptoms, it is dependent on patients’ understanding as well as willingness to endorse these items. Alternatively, the modified Raskin scale allowed for a more refined assessment of these vegetative and behavioral symptoms as raters were not reliant on self-report but rather able to identify and objectively code for these behaviors.

The current study was the first to assess the ability to identify depressive symptoms over a video-supported platform with a screening tool that does not depend upon self-report or a structured interview. Thus, the study adds to the existing literature by providing a behavioral-rated measure utilized by healthcare professionals to detect depressive symptoms. By accurately and quickly identifying depressive symptomatology through a video-supported platform, we would be able to offer better treatment of depressive symptoms and have positive long-term health outcomes for the chronically ill population. The results of this study show that the modified Raskin scale was a reliable assessment of symptoms that may have predictive but not concurrent validity. Further studies focusing at the item level may improve both the validity of the scale and reduce the rater burden. Additionally, when examining the scale validity, studies also should examine the scale’s ability to identify depressive symptomatology in different populations (i.e., older adults and ethnic, racial minorities) for further utility.

There were a number of limitations that limit conclusions that can be drawn from this study. First, sample size limited power to detect significant relationships. Only twenty-five individuals were included in the analysis. In addition, because we included only long-term HPN users and excluded caregivers, we potentially restricted the range of symptomatology. Another limitation was the psychometric properties of the modified Raskin scale. The modified Raskin appeared to have poor specificity, perhaps an artifact left over from the original scale that potentially influenced the results. Both the modified Raskin scale and PHQ-9 are limited in detecting and differentiating between grief and reactive depression, which are common within the HPN population. Thus, this inability to differentiate between reactive depression, grief, and “true” depression could influence and limit how the PHQ-9 and modified Raskin scale’s ability to provide an accurate presentation of depressive symptomatology within this population. Finally, rating was compromised by the poor video quality, a limitation that made rating symptoms to be difficult. Since telemedicine relies partly on the use of a video-supported platform, the need for a clear, steady picture and audio is essential. Healthcare professionals interested in using a video-supported platform will continuously face various technical complications (e.g., poor video quality) that can cause difficulties in diagnosis and treatment of their patients. The PHQ-9 and the modified Raskin scale both dispA critical and final limitation of this study was the use of the PHQ-9 as the criterion measure. The PHQ-9 is a measure of depressive symptomatology and cannot be used to replace a clinical interview. Thus, the results reported here are ambiguous. Without a true criterion measure, it is impossible to know whether the PHQ-9 or the modified Raskin is a more accurate assessment of depressive symptomatology.

Conclusion

Our ability to identify depressive symptoms over a video-supported platform requires ongoing investigation, particularly in examining the psychometric properties of the modified Raskin scale. Investigators should examine the modified scale to improve and increase internal consistency and criterion-related validity. With a measure that could aid in early identification and monitoring of depressive symptoms, clinicians along with untrained healthcare professionals may make better use of technology as a mode of treatment and increase the positive long-term health outcomes for the chronically ill population, who may depend upon the use of telemedicine.

Acknowledgments

We are grateful for the ongoing expertise and support from Dedrick Hooper, BA; Chang-Ming (Jeremy) Ko, MS, MA; Andrea Bevan, MA; Christina Khou, MA; Donna Maccan Yadrich, BS, MPA, and for the research team members who make these visits possible. The authors extend their appreciation to all patients and family members who participated in these studies for their time and evaluations of the iPad mobile healthcare visits and to the Oley Foundation, for continued advocacy of those managing lifelong complex HPN home care.

Funding Sources

The project is supported by the National Institute of Biomedical Imaging and Bioengineering (R01 EB015911), Carol Smith, Principal investigator. In addition, this study is partially supported by a Trail Blazer Award (awarded to Dr. Smith) from Frontiers: The Heartland Institute for Clinical and Translational Research, University of Kansas Medical Center (NIH U54 RR031295). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Biomedical Imaging and Bioengineering or the National Institutes of Health.

Contributor Information

Natasia Adams, Department of Psychology, University of Kansas

Nancy Hamilton, Associate Professor, Department of Psychology, University of Kansas

Eve-Lynn Nelson, Professor, Pediatrics, School of Medicine, Director, KU Center for Telemedicine, University of Kansas Medical Center

Carol E. Smith, Professor, School of Nursing and Preventive Medicine & Public Health, University of Kansas Medical Center.

References

  1. Baron JS, Hirani S, Newman SP. A randomized, controlled trial of the effects of a mobile telehealth intervention on clinical and patient-reported outcomes in people with poorly controlled diabetes. Journal of Telemedicine and Telecare. 2016 doi: 10.1177/1357633X16631628. [DOI] [PubMed] [Google Scholar]
  2. Baxter JP, Fayers PM, McKinlay AW. A review of the quality of life of adult patients treated with long-term parenteral nutrition. Clinical Nutrition (Edinburgh, Scotland) 2006;25(4):543–53. doi: 10.1016/j.clnu.2006.05.003. [DOI] [PubMed] [Google Scholar]
  3. Chambers A, Hennessy E, Powell-Tuck J. Longitudinal trends in quality of life after starting home parenteral nutrition: a randomised controlled study of telemedicine. Clinical Nutrition (Edinburgh, Scotland) 2006;25(3):505–514. doi: 10.1016/j.clnu.2006.01.001. [DOI] [PubMed] [Google Scholar]
  4. Chambers A, Powell-Tuck J. Determinants of quality of life in home parenteral nutrition. Current Opinion in Clinical Nutrition and Metabolic Care. 2007;10(3):318–323. doi: 10.1097/MCO.0b013e3280ebb438. [DOI] [PubMed] [Google Scholar]
  5. Chavooshi B, Mohammadkhani P, Dolatshahee B. Telemedicine vs. in-person delivery of intensive short-term dynamic psychotherapy for patients with medically unexplained pain: a 12-month randomized, controlled trial. Journal of Telemedicine and Telecare. 2016 doi: 10.1177/1357633X15627382. Advanced online publication. [DOI] [PubMed] [Google Scholar]
  6. Choi NG, Marti CN, Bruce ML, Hegel MT, Wilson NL, Kunik ME. Six-month postintervention depression and disability outcomes of in-home telehealth problem-solving therapy for depressed, low-income homebound older adults. Depress Anxiety. 2014;31(8):653–661. doi: 10.1002/da.22242. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Cummins N, Scherer S, Krajewski J, Schnieder S, Epps J, Quatieri TF. A review of depression and sucide risk assessment using speech analysis. Speech Communication. 2015;71:10–49. [Google Scholar]
  8. De Las Cuevas C, Arredondo MT, Cabrera MF, Sulzenbacher H, Meise U. Randomized clinical trial of telepsychiatry through videoconference versus face-to-face conventional psychiatric treatment. Telemedicine Journal and E-Health : The Official Journal of the American Telemedicine Association. 2006;12(3):341–350. doi: 10.1089/tmj.2006.12.341. [DOI] [PubMed] [Google Scholar]
  9. Fish J. In: Raskin Rating Scale. Kreutzer JS, DeLuca J, Caplan B, editors. New York, NY: Springer; 2011. [Google Scholar]
  10. Foley GN, Gentile JP. Nonverbal communication in psychotherapy. Psychiatry. 2010;7(6):38–44. [PMC free article] [PubMed] [Google Scholar]
  11. Hallgren KA. Computing inter-rater reliability for observational data: an overview and tutorial. Tutorials in Quantitative Methods for Psychology. 2012;8(1):23–34. doi: 10.20982/tqmp.08.1.p023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Howard L. Home parenteral nutrition: survival, cost, and quality of life. Gastroenterology. 2006;130(2 Suppl 1):S52–S59. doi: 10.1053/j.gastro.2005.09.065. [DOI] [PubMed] [Google Scholar]
  13. Howard LJ. Length of life and quality of life on home parenteral nutrition. Journal of Parenteral and Enteral Nutrition. 2002;26(5):S55–S859. doi: 10.1177/014860710202600514. [DOI] [PubMed] [Google Scholar]
  14. Huisman-de Waal G, Naber T, Schoonhoven L, Persoon A, Sauerwein H, van Achterberg T. Problems experienced by patients receiving parenteral nutrition at home: results of an open interview study. Journal of Parental and Enteral Nutrition. 2006;30(3):215–221. doi: 10.1177/0148607106030003215. [DOI] [PubMed] [Google Scholar]
  15. Huisman-de Waal G, Schoonhoven L, Jansen J, Wanten G, van Achterberg T. The impact of home parenteral nutrition on daily life-a review. Clinical Nutrition (Edinburgh, Scotland) 2007;26(3):275–288. doi: 10.1016/j.clnu.2006.10.002. [DOI] [PubMed] [Google Scholar]
  16. Ingram R. Self-focused attention in clinical disorders: Review and a conceptual model. Psychological Bullentin. 1990;107(2):156–176. doi: 10.1037/0033-2909.107.2.156. [DOI] [PubMed] [Google Scholar]
  17. Khatri N, Marziali E, Tchernikov I, Shepherd N. Comparing telehealth-based and clinic-based group cognitive behavioral therapy for adults with depression and anxiety: a pilot study. Clinical Interventions in Aging. 2014;9:276–279. doi: 10.2147/CIA.S57832. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kidholm K, Ekeland AG, Jensen LK, Rasmussen J, Pedersen CD, Bowes A, Bech M. A model for assessment of telemedicine applications: mast. International Journal of Technology Assessment in Health Care. 2012;28(1):44–51. doi: 10.1017/S0266462311000638. [DOI] [PubMed] [Google Scholar]
  19. Kim H, Spaulding R, Werkowitch M, Yadrich D, Piamjariyakul U, Gilroy R, Smith CE. Costs of Multidisciplinary parenteral nutrition care provided at a distance via mobile tablets. Journal of Parental and Enteral Nutrition. 2014;38(2):50S–57S. doi: 10.1016/j.drugalcdep.2008.02.002.A. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kirby DF. Advances in nutrition. Gastroenterology & Hepatology. 2012;8(1):39–41. [PMC free article] [PubMed] [Google Scholar]
  21. Koo TK, Li MY. A guideline of selecting and reporting intraclass correlation coefficients for reliability research. Journal of Chiropractic Medicine. 2016;15(2):155–163. doi: 10.1016/j.jcm.2016.02.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Kroenke K, Spitzer RL, Williams JB. The PHQ-9: a validity of a brief depression severity measure. Journal of General Internal Medicine. 2001;16(9):606–613. doi: 10.1046/j.1525-1497.2001.016009606.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Metzger LC. Education materials for home nutrition support consumers. Nutrition in Clinical Practice : Official Publication of the American Society for Parenteral and Enteral Nutrition. 2010;25(5):451–470. doi: 10.1177/0884533610379608. [DOI] [PubMed] [Google Scholar]
  24. Mochari-Greenberger H, Vue L, Luka A, Peters A, Pande RL. A tele-behavioral health intervention to reduce depression, anxiety, and stress and improve diabetes self-management. Telemedicine journal and e-health. 2016;22(8):624–630. doi: 10.1089/tmj.2015.0231. [DOI] [PubMed] [Google Scholar]
  25. Mor N, Winquist J. Self-focused attention and negative affect: A meta-analysis. Psychological Bulletin. 2002;128(4):638–662. doi: 10.1037/0033-2909.128.4.638. [DOI] [PubMed] [Google Scholar]
  26. Nezu AM, Ronan GF, Meadows EA, McClure KS, editors. Practitioner’s Guide to Empirically Based Measures of Depression AABT Clinical Assessment Series. New York, NY: Kluwer Academy; 2000. [Google Scholar]
  27. O’Reilly R, Bishop J, Maddox K, Hutchinson L, Fisman M, Takhar J. Is telepsychiatry equivalent to face-to-face psychiatry? Results from a randomized controlled equivalence trial. Psychiatric Services. 2007;58(6):836–843. doi: 10.1176/appi.ps.58.6.836. [DOI] [PubMed] [Google Scholar]
  28. Persoon A, Huisman-de Waal G, Naber TA, Schoonhoven L, Tas T, Sauerwein H, van Achterberg T. Impact of long-term HPN on daily life in adults. Clinical Nutrition (Edinburgh, Scotland) 2005;24(2):304–313. doi: 10.1016/j.clnu.2004.12.009. [DOI] [PubMed] [Google Scholar]
  29. Pyszczynski T, Hamilton JC, Hering FH, Greenberg J. Depression, self-focused attention, and the negative memory. Journal of Personality and Social Psychology. 1989;57(2):351–357. [Google Scholar]
  30. Raskin A, Schulterbrandt J, Reatig N, McKeon J. Differential response to chlorpromazine, imipramine and placebo. A study of sub-groups of hospitalised depressed patients. Archives of General Psychiatry. 1970;23(2):164–173. doi: 10.1001/archpsyc.1970.01750020068009. [DOI] [PubMed] [Google Scholar]
  31. Saqui O, Chang A, McGongle S, Purdy B, Fairholm L, Baun M, Yeung M, Allard J. Telehealth videoconferencing : improving home parenteral Nutrition. Journal of Parental and Enteral Nutrition. 2007;31(3):234–239. doi: 10.1177/0148607107031003234. [DOI] [PubMed] [Google Scholar]
  32. Salisbury C, O’Cathain A, Edwards L, Thomas C, Gaunt D, Hollinghurst S, Montgomery AA. Effectiveness of an integrated telehealth service for patients with depression: a pragmatic randomized controlled trial of a complex intervention. Lancet Psychiatry. 2016;3(6):515–525. doi: 10.1016/S2215-0366(16)00083-3. [DOI] [PubMed] [Google Scholar]
  33. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychological Bulletin. 1979;86(2):420–428. doi: 10.1037//0033-2909.86.2.420. [DOI] [PubMed] [Google Scholar]
  34. Smith CE, Spaulding R, Piamjariyakul U, Werkowitch M, Yadrich D, Hooper D, Gilroy R. mHealth clinic appointment PC tablet:implementation, challenges and solutions. Journal of Mobile Technology in Medicine. 2015;4(2):21–32. doi: 10.1016/j.ygyno.2014.12.035. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Staun M, Pironi L, Bozzetti F, Baxter J, Forbes A, Joly F, Van Gossum A. ESPEN guidelines on parenteral nutrition: home parenteral nutrition (HPN) in adult patients. Clinical Nutrition (Edinburgh, Scotland) 2009;28(4):467–479. doi: 10.1016/j.clnu.2009.04.001. [DOI] [PubMed] [Google Scholar]
  36. Winkler MF. Quality of life in adult home parenteral nutrition patients. Journal of Parenteral and Enteral Nutrition. 2005;29(3):162–170. doi: 10.1177/0148607105029003162. [DOI] [PubMed] [Google Scholar]
  37. Winkler M, DiMaria-Ghalili R, Guenter P, Resnick HE, Robinson L, Lyman B, Steiger E. Characteristics of a cohort of home parenteral nutrition patients at the time of enrollment in the sustain registry. Journal of Parenteral and Enteral Nutrition. 2015;40(8):1140–1149. doi: 10.1177/0148607115586575. [DOI] [PubMed] [Google Scholar]
  38. Winkler M, Guenter P. Long-term home parenteral nutrition: it takes an interdisciplinary approach. Journal of Infusion Nursing : The Official Publication of the Infusion Nurses Society. 2014;37(5):389–395. doi: 10.1097/NAN.0000000000000068. [DOI] [PubMed] [Google Scholar]
  39. Winkler MF, Ross VM, Piamjariyakul U, Gajewski B, Smith CE. Technology dependence in home care: impact on patients and their family caregivers. Nutrition in Clinical Practice. 2006;21(6):544–556. doi: 10.1177/0115426506021006544. [DOI] [PubMed] [Google Scholar]
  40. Winkler MF, Smith CE. Clinical, social, and economic impacts of home parenteral nutrition dependence in short bowel syndrome. Journal of Parenteral and Enteral Nutrition. 2014;38(1 Suppl):32S–37S. doi: 10.1177/0148607113517717. [DOI] [PubMed] [Google Scholar]
  41. Zhao VM, Griffith DP, Blumberg HM, Dave NJ, Battey CH, McNally TA, Ziegler TR. Characterization of post hospital infections in adults requiring home parenteral nutrition. Nutrition. 2013;29(1):52–59. doi: 10.1016/j.nut.2012.03.010. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES