The Patient Health Questionnaire depression screener in spinal cord injury

Abstract

Context

Although depression is not inevitable following spinal cord injury/dysfunction (SCI/D), it can have a negative impact on rehabilitation. Evidence-based assessment of depression utilizing self-report instruments, such as the Patient Health Questionnaire-9 (PHQ-9), is considered good clinical practice. Although the PHQ-9 has been studied in individuals with SCI/D, little is known about the clinical utility of the Patient Health Questionnaire-2 (PHQ-2). Traditional cutoff scores for the PHQ-2 were examined to explore their operating characteristics as related to PHQ-9 results.

Methods

Archival data were collected for 116 Veterans with SCI/D who completed the PHQ-2 and PHQ-9 as one component of their routine, comprehensive SCI annual evaluation at a Veterans Affairs Medical Center. Logistic regressions were performed to determine the impact of different cutoff scores for the PHQ-2 on the likelihood that participants would endorse clinically significant levels of depressive symptoms on the PHQ-9 (≥10).

Results

Using a cutoff score of 3 or greater correctly classified 94.8% of the cases, outperforming the other cutoff scores. A cutoff score of 3 or greater had a sensitivity of 83.3% and specificity of 97.8%, and yielded a positive predictive value of 90.9% and a negative predictive value of 95.7%.

Conclusion

The PHQ-2 shows promise as a clinically useful screener in the community-residing SCI/D population. Findings regarding the presence of suicidal ideation emphasize the importance of routine screening for depressive symptomatology in the SCI/D population. Future research should investigate the role of the PHQ-2 in clinical decision-making and treatment monitoring.

Introduction

Although variability exists in psychological adjustment following spinal cord injury/dysfunction (SCI/D), depression is the most common form of psychological distress following SCI/D and can have a negative impact on rehabilitation.¹ Specifically, depression in individuals with SCI/D is associated with fewer hours out of bed, fewer days out of the house, less engagement in productive activities, and less planned exercise.^2,3 Estimates of the prevalence of Major Depressive Disorder in individuals with SCI/D range from about 11% or 12% to 21% 1 year post-injury,^4–6 and from about 10% to 18% 5 years post-injury.^4,6 In community-residing individuals with SCI/D, about 23% meet the criteria for Major Depressive Disorder.⁷ In comparison, the lifetime prevalence of Major Depressive Disorder in the general U.S. population is 16.6%.⁸ Considering the prevalence and implications of depression in individuals with SCI/D, best practice guidelines recommend evidence-based depression assessment; however, a disconnect remains between best evidence and current practice.⁹

A systematic review of depression measures used in SCI/D research suggested that no one instrument is psychometrically superior. Instead, it was concluded that selection of a depression measure should be made based on other factors, including feasibility, acceptability to patients, ease of administration and scoring, and the ability of the measure to serve additional purposes, such as monitoring therapy outcomes.¹⁰ Indeed, one of the primary barriers to efficient and effective screening, diagnosis, and treatment of depressive disorders is the fact that many screening measures are considered too long (i.e. 20 plus items) to be practical in routine clinical practice. An additional barrier is that most depression screening measures do not relate to diagnostic criteria for Major Depressive Disorder, requiring clinicians to conduct a separate diagnostic evaluation.⁵

The Patient Health Questionnaire-9 (PHQ-9) addresses these barriers for several reasons.¹¹ Consisting of only nine items, the PHQ-9 is more acceptable to patients and easier for clinicians to use routinely. Importantly, the nine items parallel the diagnostic criteria for Major Depressive Disorder, making it both a depression severity measure and a diagnostic instrument. Moreover, the PHQ-9 has been validated in an acute SCI/D rehabilitation sample.¹² Despite research demonstrating the PHQ-9′s utility, it has not been adopted as the standardized approach to depression assessment during acute SCI/D rehabilitation or at annual evaluations.⁹ This may be due to its length,¹³ or because of the inclusion of an item assessing suicidal ideation. Specifically, it is not uncommon for nurses and primary care physicians to complete their education without having received adequate training in suicide assessment.^14,15 This lack of training has the potential to make health care professionals feel unprepared to complete a risk assessment if a patient endorses suicidal ideation on a self-report questionnaire.

Research has provided empirical support for use of the PHQ-2, an even shorter screener, in a VA urgent care clinic sample,¹⁶ and in U.S. primary care and obstetrics-gynecology clinic samples.¹⁷ Lowe et al. demonstrated that the diagnostic accuracy of the PHQ-2 is comparable with longer measures and that it is sensitive to change and accurately reflects different courses of depression.¹⁸ Since its development, the PHQ-2 has been used in primary care samples across various countries as well as in several other medical populations (Table 1). Although research has been conducted on use of the PHQ-2 in an acute SCI/D rehabilitation population,¹² additional evidence is needed to justify the clinical utility of this depression screener in the community-residing SCI/D population. The present study aims to address this evidence gap using data collected from Veterans with SCI/D who presented for their annual evaluation. It was hypothesized that the results of this study would establish the clinical utility of the PHQ-2 in community-residing individuals with SCI, thereby enabling providers to identify more efficiently which individuals may benefit from additional assessment and treatment.

Table 1.

PHQ-2 operating characteristics in medical samples.

Population	Authors/year	PHQ-2 Cutoff score	PHQ-2 Operating characteristics	PHQ-9 Cutoff score	PHQ-9 Operating characteristics
Primary Care	Kroenke et al. (2003)17	≥3	Sensitivity = 83%; Specificity = 90%; PPV = 38.4%; AUC = 0.93	Not reported	AUC = 0.95
	Corson et al. (2004)19	≥3	Sensitivity = 97%; Specificity = 91%; AUC = 0.94	Not reported	Not reported
	Chen et al. (2009)20	≥3	Sensitivity = 84%; Specificity = 90%; AUC = 0.92	≥9	Sensitivity = 86%; Specificity = 85%; AUC = 0.92
	Osorio et al. (2009)21	≥3	Sensitivity = 97%; Specificity = 88%; PPV = 81%; NPV = 98%; AUC = 0.97	≥10	Sensitivity = 100%; Specificity = 98%; PPV = 97%; NPV = 100%; AUC = 0.998
	Arroll et al. (2009)22	≥2	Sensitivity = 86%; Specificity = 78%	≥10	Sensitivity = 74%; Specificity = 91%;
	Phelan et al. (2010)23	≥2	Sensitivity = 75%; Specificity = 67%; AUC = 0.81	≥9	Sensitivity = 88%; Specificity = 80%; AUC = 0.87
	Zuithoff et al. (2010)24	≥2	Sensitivity = 81%; Specificity = 76%; PPV = 34%; NPV = 96%; AUC = 0.83	≥6	Sensitivity = 82%; Specificity = 82%; PPV = 41%; NPV = 97%; AUC = 0.87
	Liu et al. (2011)25	≥2	Sensitivity = 88%; Specificity = 82%; AUC = 0.9	≥10	Sensitivity = 86%; Specificity = 93.9%; AUC = 0.96
	Inagaki et al. (2013)26	≥2	Sensitivity = 77%; Specificity = 95%; PPV = 54%; NPV = 98%; AUC = 0.95	≥4	Sensitivity = 86%; Specificity = 85%; PPV = 32%; NPV = 99%; AUC = 0.93
	Thapar et al. (2014)27	≥3	Sensitivity = 72.1%; Specificity = 82.1%; PPV = 53.5%; NPV = 91.2%; AUC = 0.87	≥11	Sensitivity = 83.6%; Specificity = 83%; PPV = 58.4%; NPV = 94.7%; AUC = 0.9
	Bhana et al. (2015)28	≥2	Sensitivity = 60%; Specificity = 84%; AUC = 0.76	≥9	Sensitivity = 49%; Specificity = 94%; AUC = 0.85
	Carey et al. (2015)29	≥3	Sensitivity = 91%; Specificity = 78%; PPV = 41%; NPV = 98%; AUC = 0.92	≥10	Not reported
	Hanlon et al. (2015)30	≥1	Sensitivity = 83.3%; Specificity = 60.8%; PPV = 11.7%; NPV = 98.3%; AUC = 0.78	≥5	Sensitivity = 83.3%; Specificity = 74.7%; PPV = 17.1%; NPV = 98.6%; AUC = 0.85
	Suzuki et al. (2015)31	≥3	Sensitivity = 76%; Specificity = 82%; Overall Accuracy = 81%; PPV = 27%; NPV = 98%; AUC =.845	≥11	Sensitivity = 76%; Specificity = 81%; Overall Accuracy = 81%; PPV = 26%; NPV = 97%; AUC = 0.88
Stroke	de Man-van Ginkel, Gooskens, et al. (2012)32	≥2	Sensitivity = 100%; Specificity = 77%; PPV = 38%; NPV = 100%	≥10	Sensitivity = 100%; Specificity = 86%; PPV = 50%; NPV = 100%
	de Man-van Ginkel, Hafsteinsdottir, et al. (2012)33	≥2	Sensitivity = 75%; Specificity = 76%; AUC = 0.82	≥10	Sensitivity = 80%; Specificity = 78%; AUC = 0.87
	Turner et al. (2012)34	≥1	Sensitivity = 77%; Specificity = 63%; AUC = 0.82	≥6	Sensitivity = 85%; Specificity = 63%; AUC = 0.82
Cardiovascular	McManus et al. (2005)35	≥3	Sensitivity = 39%; Specificity = 92%; AUC = 0.84	≥10	Sensitivity = 54%; Specificity = 90%; AUC = 0.86
	Thombs et al. (2008)36	≥2	Sensitivity = 82%; Specificity = 79%; PPV = 52%; NPV = 94%; AUC = 0.84	≥6	Sensitivity = 83%; Specificity = 76%; PPV = 50%; NPV = 94%; AUC = 0.86
	Wang et al. (2015)37	≥3	Sensitivity = 85.7%; Specificity = 69.2%; PPV = 57.1%; NPV = 93.6%; AUC = 0.806	≥10	Sensitivity = 87.1%; Specificity = 83.5%; PPV = 58.7%; NPV = 95.6%; AUC = 0.877
HIV/AIDS	Monahan et al. (2008)38	≥3	Sensitivity = 91%; Specificity = 77%; PPV = 36.9; AUC = 0.91	Not reported	Not reported
Cognitive Impairment	Boyle et al. (2011)39	≥3	Sensitivity = 78%; Specificity = 71%; AUC = 0.81	≥10	Sensitivity = 89%; Specificity = 71%; AUC = 0.85
Epilepsy	Fiest et al. (2014)40	≥2	Sensitivity = 42.3%; Specificity = 87.3%; PPV = 35.5%; NPV = 90.1%; AUC = 0.75	≥9	Sensitivity = 82.6%; Specificity = 82.2%; PPV = 42.2%; NPC = 96.8%; AUC = 0.88
Migraine	Seo et al. (2015)41	≥2	Sensitivity = 66.7%; Specificity = 90.3%; PPV = 74.3%; NPV = 86.6%; AUC = 0.785	≥7	Sensitivity = 79.5%; Specificity = 81.7%; PPV = 64.6%; NPV = 90.5%; AUC = 0.806
Multiple Sclerosis	Amtmann et al. (2015)42	≥2	Sensitivity = 70.8%; Specificity = 81%	≥10	Sensitivity = 93.8%; Specificity = 61.2%
	Patten et al. (2015)43	≥3	Sensitivity = 80%; Specificity = 93%; PPV = 64%; NPV = 36%; AUC = 0.943	≥11	Sensitivity = 95%; Specificity = 88.3%; PPV = 55.9%; NPV = 44.1%; AUC = 0.952

Methods

Participants and procedure

Archival data were collected from a convenience outpatient sample of Veterans with SCI/D who presented to a Veterans Affairs Medical Center during 2012 for their routine, comprehensive SCI/D annual evaluation. These evaluations were completed by an interdisciplinary team of SCI/D providers. Of 144 Veterans assigned to be assessed by the senior author of this manuscript, 116 (80.6%) completed the PHQ-2 and PHQ-9 among other brief psychological screeners. Participants completed the measures in written or oral format; electronic medical records were reviewed to identify demographic and injury characteristics. This study was approved and monitored for compliance with ethical research practices by an Internal Review Board (H-30172) and the VA Research and Development Committee (12E01.H).

Measures

Patient Health Questionnaire-9 (PHQ-9)¹¹

The PHQ-9 is the nine item depression module from the full Patient Health Questionnaire.⁴⁴ It consists of the nine criteria upon which diagnoses of depressive disorders are based according to the Diagnostic and Statistical Manual of Mental Disorders. The PHQ-9 can assist clinicians in diagnosing depressive disorders as well as determining depressive symptom severity. Participants are instructed to indicate how often over the past 2 weeks they have been bothered by nine symptoms of depression with response options ranging from ‘not at all’ (0) to ‘nearly every day’ (3). Responses are summed to obtain a total score that can range from 0 to 27; scores of 5, 10, 15, and 20 serve as cutoff scores that indicate mild, moderate, moderately severe, and severe depression.¹¹ The internal consistency of the data was excellent in both a primary care sample of 3,000 (α = 0.89) and in an obstetrics-gynecology sample of 3,000 (α = 0.86), and it was demonstrated that the PHQ-9 discriminated well between individuals with and without depression (AUC = 0.95). It was concluded that scoring the PHQ-9 as a continuous measure and using a cutoff score of 10 or higher results in the most accurate predictions of Major Depressive Disorder when compared with independent diagnoses made by a mental health provider.¹¹ In the present study, the data had strong internal consistency, Cronbach's alpha = 0.95.

Patient Health Questionnaire-2 (PHQ-2)¹⁷

The PHQ-2 consists of the first two items from the PHQ-9. In a primary care and obstetrics-gynecology clinic sample of 6,000 individuals, the PHQ-2 discriminated well between individuals with and without depression (AUC = 0.93).¹⁷ In the present study, the data had strong internal consistency, Cronbach's alpha = 0.91.

Data analysis

Descriptive statistics were used to detail demographic and injury characteristics. Logistic regressions were performed to determine the impact of different cutoff scores for the PHQ-2 on the likelihood that participants would endorse clinically significant levels of depressive symptoms on the PHQ-9 (≥10). Receiver operating characteristic (ROC) curve analysis was also conducted.

Results

Demographic and injury characteristics of the 116 participants are reported in Table 2. The sample was mostly middle-aged, married, White men with incomplete traumatic injuries. These characteristics are comparable to the 2007 fiscal year national data reported on Veterans with SCI/D using VA services.⁴⁵

Table 2.

Demographic and injury characteristics.

Characteristic	Study sample mean (SD) or %
Age (Years)	56.0 (12.4)
Time Since SCI/D (Years)	18.1 (13.3)
Sex
Male	96.6
Female	3.4
Race/ethnicity
White	56.0
Black	33.6
Hispanic/Latino	7.8
Native American	1.7
Other	.9
Marital Status
Married	49.1
Divorced	33.6
Single	14.7
Separated	2.6
Injury Characteristics
Tetraplegia (AIS A, B, C)	24.2
Paraplegia (AIS A, B, C)	31.7
AIS D	38.8
Traumatic Injury
Yes	81.0
No	19.0

Using a PHQ-9 cutoff score of 10 or higher as a proxy for diagnosis, the estimated prevalence of Major Depressive Disorder in this sample was 20.7%. In this outpatient, non-psychiatric sample of Veterans with SCI/D, 12% endorsed experiencing suicidal ideation several days or more during the previous 2 weeks.

The average PHQ-2 score was 1.17 (SD = 1.948). The average PHQ-9 score was 5.23 (SD = 7.451). The operating characteristics of the PHQ-2 at various cutoff scores are reported in Table 3. Using a cutoff score of 3 or greater outperformed other cutoff scores, correctly classifying 94.8% of the cases. This cutoff score had a sensitivity of 83.3% and a specificity of 97.8%, yielding a positive predictive value of 90.9% and a negative predictive value of 95.7%. Additionally, receiver operating characteristic (ROC) analysis revealed an AUC value of 0.979 for the PHQ-2, suggesting that the PHQ-2 possesses excellent diagnostic accuracy in the present study's sample of community-residing individuals with SCI/D.⁴⁶

Table 3.

PHQ-2 operating characteristics.

PHQ-2 Cutoff score	Percentage accuracy in classification	Sensitivity	Specificity	Positive predictive value	Negative predictive value
1	84.5%	100%	80.4%	57.1%	100%
2	92.2%	95.8%	91.3%	74.2%	98.8%
3	94.8%	83.3%	97.8%	90.9%	95.7%
4	93.1%	70.8%	98.9%	94.4%	92.9%
5	89.7%	54.2%	98.9%	92.9%	89.2%

Discussion

The present study extended previous research findings by demonstrating that the PHQ-2 shows promise as a clinically useful depression screener in the community-residing SCI/D population. Specifically, a cutoff score of 3 or greater outperformed other scores, but the cutoff score of 2 or greater could be used if the goal is to maximize sensitivity at the expense of specificity. Regardless of the cutoff score selected, the PHQ-2 was shown to possess excellent diagnostic accuracy. Additionally, the estimated prevalence of Major Depressive Disorder in the sample was 20.7%. This is consistent with previous research indicating that in community-residing individuals with SCI/D, about 23% met the criteria for Major Depressive Disorder, also using the PHQ-9 as a proxy for diagnosis.⁷

The results of the present study provide support for the clinical utility of the PHQ-2. Clinical utility is conceptualized as consisting of many factors relevant to clinical practice, including how easily an instrument or intervention can be learned and applied by different practitioners, how it fits in with the idiosyncrasies of a setting, and its compatibility with the values of the culture in which it is applied.⁴⁷ According to Smart's multidimensional model, instruments or interventions that possess clinical utility are appropriate, accessible, practical, and acceptable. In other words, they must be effective (i.e. supported by formal evidence) and relevant (i.e. important for clinical decision-making), easily accessed both in terms of cost and availability, functional in and suitable for the specific environment, and acceptable to clinicians and clients.⁴⁸ Specific to screening for depression, Kroenke asserted that clinically useful measures are brief, easily administered, multipurpose (i.e. can be used for screening, severity assessment, probable diagnosis, and treatment monitoring), free, and easy to score (i.e. single summative score without complex procedures).¹³ Kroenke also emphasized that determining an optimal cutoff score and showing sensitivity to change are essential aspects of clinical utility. The PHQ-2 meets several of the requirements for depression screening measures recommended by Kroenke (i.e. brief, easily administered, free, easy to score).¹³ Likewise, based on Smart's multidimensional model of clinical utility and specific to the circumstances of the present study, the PHQ-2 was effective (i.e. supported by formal evidence), easily accessed both in terms of cost and availability, and functional in and suitable for the specific environment.⁴⁸

It is important to note both the strengths and the limitations of the present study. Study measures were administered within the context of routine care, suggesting that the findings are more likely to generalize to other real-life clinical settings. Moreover, the measures were administered independently as opposed to taking the first two items from the PHQ-9 to constitute the PHQ-2. Although a convenience sample, participant demographics were similar to the national VA SCI/D sample.⁴⁵ However, the sample was of limited diversity, consisting entirely of Veterans, most of whom were male and many of whom had incomplete traumatic injuries. Thus, further research is needed to determine if the results of the present study may generalize to the broader SCI/D population. Another potential limitation of the present study is that the measures were administered by behavioral health providers, but because of the ease of administration and scoring, the PHQ-2 can be and is routinely administered by other health care providers, such as nurses.^32,33 This is especially valuable in SCI/D care settings that may not have a full-time behavioral health provider on staff to conduct depression screenings.

Future research should investigate the ways in which the PHQ-2 can address other aspects of clinical utility, such as relevance in clinical decision-making and treatment monitoring. For example, the PHQ-2 could be administered at intervals throughout treatment of depression in a community-residing SCI/D population to determine its sensitivity to change over time and its relevance in clinical decision-making. Clients and clinicians also could be surveyed about their perceptions of the PHQ-2 in order to study an additional component of Smart's model, which is acceptability to clinicians and clients.⁴⁸ Overall, the results of this study demonstrate the clinical utility of the PHQ-2 in community-residing individuals with SCI/D as a screening measure that reflects the best available evidence while remaining sensitive to the needs of the clinical environment.

Appendix

The Patient Health Questionnaire-9¹¹

Response Options:

Not at all (0); Several days (1); More than half the days (2); Nearly every day (3)

Over the last 2 weeks, how often have you been bothered by any of the following problems?

• 1. Little interest or pleasure in doing things
• 2. Feeling down, depressed, or hopeless
• 3. Trouble falling or staying asleep, or sleeping too much
• 4. Feeling tired or having little energy
• 5. Poor appetite or overeating
• 6. Feeling bad about yourself – or that you are a failure or have let yourself or your family down
• 7. Trouble concentrating on things, such as reading the newspaper or watching television
• 8. Moving or speaking so slowly that other people could have noticed? Or the opposite – being so fidgety or restless that you have been moving around a lot more than usual
• 9. Thoughts that you would be better off dead or of hurting yourself in some way

Disclaimer statements

Contributor None.

Funding None.

Conflicts of interest The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs, the US government, Baylor College of Medicine, or Texas A&M College of Medicine. None of these bodies played a role in study design; in the collection, analysis and interpretation of data; in the writing of the report; or in the decision to submit the article for publication.

Ethics approval None.

ORCID

Joseph Mignogna http://orcid.org/0000-0001-8618-2607

Aimee J. Christie http://orcid.org/0000-0003-2558-966X