Mental Health Screening of Older Adults in Primary Care

Abstract

In an effort to document mental health outreach in our primary care clinic, 316 veterans (mean age 72) not currently in psychiatric treatment were screened for multiple mental health symptoms. Depressed mood was reported by 18% of the sample, insomnia by 26%, and morbid/suicidal ideation by 6.9% for at least several days during the past 2 weeks. Of those who experienced a loss over the past year (43%), 36% remained affected by the loss. Also reported were anxiety symptoms (29%) and PTSD symptoms (14%). Two-fifths (39%) of patients reported drinking alcohol in the past week, 18% more than 5 days, and 13% more than 3 drinks per sitting. Twenty-six percent of the patients reported symptoms warranting intervention; of these, only 39% accepted a treatment referral. While screening for depressed mood and alcohol use is now common in primary care, we found it useful to screen for specific symptoms of depression (including insomnia and suicidal ideation), persisting grief reactions, anxiety, and PTSD in this setting. Further research is necessary to determine factors that underlie some patients’ refusal to accept mental health treatment.

Older adults who experience psychiatric disorders rarely seek care from mental health specialists; rather, they tend to seek help from their general medical physician (Gallo, Rabins, & Illife, 1997). The primary health care clinic is therefore an important setting for the detection of mental health symptoms and subsequent treatment. Yet, older adults often present with a complex combination of medical, neurological, and psychological symptoms that increase the complexity of identifying and treating psychiatric difficulties (Jeste et al., 1999). As such, primary care providers have been noted to under-diagnose mental health problems and to under-treat such problems when recognized (Unützer et al., 2000). Such under-diagnosis and treatment is costly as older adults with depression, anxiety, and/or alcohol abuse difficulties are more disabled by medical illness, use more health care services, and have higher rates of mortality (Jeste et al., 1999). Fortunately, effective treatments for depression (e.g., Karel & Hinrichsen, 2000), alcohol abuse (e.g., Schonfeld & Dupree; 1995), and anxiety disorders (e.g., Stanley & Beck, 2000) are available for older adults.

Screening for Depression

While only 1% to 2% of community-dwelling elders have major depressive disorder, 8% to 20% report clinically significant depressive symptoms (Gallo & Lebowitz, 1999). Rates of depression are higher in medical settings where up to 35% in primary care settings report minor depression (Gurland, Cross, & Katz, 1996). Identifying symptoms of depression among older adults is critical, considering the increased morbidity and mortality associated with depression including suicide (Conwell et al, 2000); a majority of patients who commit suicide visit their physician within a month of the suicide (Caine, Lyness, & Conwell, 1996).

Screening for Grief Reactions

Ten to twenty percent of widowers and widows experience depression during the first year following the death of a.spouse (Zisook & Shuchter, 1993). The negative psychological effects of the death of a spouse may continue for two or more years following the bereavement including increased rates of physical illness, increased alcohol consumption, and early death (Rosenzweig, Prigerson, Miller, & Reynolds, 1997).

Screening for Anxiety and Post Traumatic Stress Disorder (PTSD)

While only 5% of older adults have a diagnosable anxiety disorder (Flint, 1994), up to 21 % of older adults have reported anxiety symptoms that do not meet full diagnostic criteria (Himmelfarb & Murrell, 1984). The relative contribution of anxiety versus medical conditions to physical symptomatology can be particularly difficult to differentiate for older adults, such as in pulmonary (Kim et al., 2000) and rheumatological (O’Malley et al., 1998) conditions.

Because of the massive mobilization for World War II followed closely by Korea, nearly half of the current cohort of older men served in the military, and about half of these were exposed to combat (Spiro, Schnurr, & Aldwin, 1994). One-fifth to one-third of older veterans seen in a medical clinic may report PTSD symptoms (Blake et al., 1990), which is associated with poorer self-reported physical health (Clipp & Elder, 1996). Some World War II and Korean era combat veterans have been asymptomatic until they experience major transitions associated with late life (e.g., retirement) or a serious medical illness (Schnurr & Friedman, 1997).

Screening for Alcohol Abuse

Alcohol abuse puts elders at risk for multiple health, cognitive, psychiatric, and interpersonal problems (Substance Abuse and Mental Health Services Administration [SAMSA], 1998). Ten to fifteen percent of older adults treated in primary care clinics have problematic alcohol use (exceeding a recommended one drink: per day) (Adams, Barry, & Fleming, 1996), with older adults who have been separated, divorced, or widowed at increased risk (Bucholz, Sheline, & Helzer, 1995). Regular screening by primary care clinicians is useful as many of those at risk do not seek services for substance abuse problems on their own yet are responsive to brief interventions in medical care settings (Fleming, Barry, Manwell, Johnson, & London, 1997).

In consideration of the potential benefits of screening and then treating older medical patients for mental health problems, we describe our efforts to provide such screening of older veterans in a Department of Veterans Affairs primary care clinic. We report the extent of unrecognized or untreated mental health concerns, the feasibility of outreach screening for this population, and the response of patients to a mental health referral. Results are discussed in light of our clinical experience treating medically ill, elderly veterans with late onset mental health problems. A strength of this report is its focus on screening for multiple mental health symptoms, however. Limitations include that we did not collect data to permit the comparison of screening data and subsequent diagnostic interviews.

METHOD

Participants

Three hundred eighty-two veterans waiting for primary care appointments were approached for mental health screening; 316 completed the interview. Some patients declined to be interviewed. Other interviews were interrupted if the physician became available to see the patient.

Instruments

Demographic Information

Demographic information was obtained by asking patients to report their age, ethnicity, educational attainment, and occupational and marital status as well as current living arrangements.

Depression

Depression was assessed with the Patient Health Questionnaire of the Primary Care Evaluation of Mental Disorders - Depression Subscale (PHQ/PRIME-MD; Spitzer et al., 1999). The PHQ Depression Subscale uses a 4-point likert scale to rate nine DSM-IV symptoms of major depression over the past two weeks as: not at all, several days, more than half the days, or nearly every day. The PRIME-MD is designed for use by primary care physicians to detect the psychiatric disorders most commonly found in primary care settings. Diagnoses made by primary care physicians using the PRIME-MD versus experienced mental health professionals are comparable (Kappa = .65). In this study, the PHQ was administered in interview format to avoid potential problems with literacy or vision.

Grief Reactions

Grief was assessed by asking patients: Have you lost a close friend or family member in the past year? They were also asked: Do you find that it still affects you? Both items were rated as yes/no.

Anxiety and Post Traumatic Stress Disorder

Two questions from the PHQ were used to assess anxiety level over the last 4 weeks. The first was: How often have you felt nervous (anxious), on edge, or worried a lot about different things? The second was: How often have you become easily annoyed or irritable? These questions were rated as: not at all, several days, or more than half the days. Additional PHQ questions for anxiety were eliminated as we were not interested in evaluating specific anxiety disorders in this screening study.

Seven questions based on the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders fourth edition (APA, 1994) were used to assess Symptoms of Post Traumatic Stress Disorder about military or other trauma. These questions focused on military experiences of combat or having been involved with injured or dying servicemen, and whether veterans currently have distressing memories, dreams, acts or feelings of recurrence, psychological distress, or physiologic reactivity upon exposure to traumatic cues, in relation to military or other trauma.

Alcohol Use

Five questions adapted from the Alcohol Use Disorders Identification Test (AUDIT) (Babor, de la Fuente, Saunders, & Grant, 1992) were used to assess current alcohol abuse focusing on frequency and quantity in drinking as well as social concerns about drinking. We did not include questions focusing on alcohol dependence (e.g., “How often during the last 6 months have you needed a drink in the morning to get yourself going after a heavy drinking sessions?”) as we have found these apply to only a small number of older adults.

Procedure

Lists of potential participants were reviewed by hospital computer on the day prior to the participants’ primary care appointment to identify those patients 60 years of age and older who were not currently in mental health treatment.

Working from this list of potential participants, patients were approached while waiting for an appointment if they were expecting at least a 15-minute wait to see the physician (as assessed by the clinic clerk). The interviewer was a clinical psychology graduate student who had completed a full-time internship and three years of geriatric mental health screening experience. Patients were informed that the Medical Center’s Geriatric Mental Health Clinic was interviewing older adults to assess if they might be experiencing difficulties which could be improved by appropriate treatment. Patients were also informed that participation in the screening interview was entirely voluntary and that refusal to participate would not affect their provision of care at the Medical Center. Patients who agreed to participate were screened in a private room. Interviews were terminated if the physician became available to see the patient. Most interviews were completed in 15 minutes.

At the end of the interview, all patients were asked whether they would like a referral to the Geriatric Mental Health clinic to discuss any issues further. If more urgent risk issues were identified (e.g., suicidal thoughts), the primary care provider was notified and further evaluation by a staff psychiatrist or psychologist was conducted as indicated. Following the interview, all participants were rated by the interviewer as to whether they needed mental health services on the basis of interview responses; notation was made of whether participants agreed or declined to be contacted by a provider in our clinic. Those patients who expressed an interest in a referral were contacted by a provider in the Medical Center’s Geriatric Mental Health Clinic.

RESULTS

Sample Characteristics

Three hundred and sixteen patients completed the interview, with a mean age of 72.1 years, (SD = 6.4) and educational level of 11.7 years (SD = 3.0). Most were Caucasian (94.9%), 4.8% were African American and .3% were Hispanic. The majority of participants were married (54.5%), 18.8% were widowed, 17.8% were divorced, 5.1% were never married, and 3.5% were separated. Many patients reported living with a spouse (43.2%), while 30.2% lived alone and 26.6% had other living arrangements. The majority were retired (71.5%), 19.2% were working part time, 3.2% were working full time, and 6.1% held regular volunteer positions.

Depression

Six percent of patients reported lost interest or pleasure, and 9% more described feeling down, depressed, or hopeless more than half the days in the past two weeks (see Table 1). Almost 7% said they had thought they’d be better off dead or had thoughts of hurting themselves at least several days in the past two weeks.

TABLE 1.

Depressive Symptoms on the PRIME MD.

	Not at all	Several days	> than 1/2 the days*
Lost interest or pleasure in doing things	87.2%	6.7%	6.1%
Felt down, depressed, or hopeless	73.8%	17.3%	9.0%
Trouble falling or staying asleep	66.6%	12.2%	21.2%
Tired or had little energy	66.8%	13.7%	19.5%
Poor appetite or overeating	89.5%	4.8%	5.7%
Felt bad about self or a failure or   had let self/family down	89.2%	8.0%	2.9%
Trouble concentrating on things, such as   reading the newspaper or watching TV	87.8%	7.4%	4.8%
Moving or speaking so slowly that   other people noticed	88.5%	6.7%	4.4%
Thoughts that would be better off dead   or of hurting self in some way	92.7%	5.7%	1.2%

^*In this table the response categories “more than 1/2 the days” and “nearly every day” are collapsed in column three to “more than 1/2 the days.”

The most commonly endorsed symptoms were trouble falling asleep and feeling tired or having little energy. But such endorsement was not always associated with depressed mood or low interest/pleasure. Of those complaining of insomnia, only 24% also complained of associated mood or lost interest/pleasure; of those complaining of fatigue or anergy, only 30% also complained of the associated symptoms. Overall, 9% of patients endorsed 5 or more symptoms on the PRIME-MD (including lost interest or depressed mood) over at least several days, and 1% endorsed 5 or more symptoms nearly every day.

Grief

One hundred and thirty-six (43%) participants lost a close friend or family member in the past year; 51 (39% of those who lost someone; 16% of the total sample) report still being affected by the loss.

Anxiety and Post Traumatic Stress Disorder

Many participants reported mild anxiety symptoms. About one-fourth described feeling nervous, on edge, or worrying (21.4% several days in the past month, 7.5% more than half the days). About one-third reported feeling easily annoyed or irritable (28.5% several days, 6.3% more than half the days).

About two-fifths of the participants were either in combat or involved in some way with injured or dying servicemen (see Table 2). Fourteen percent of the total sample acknowledged having distressing memories of the war; 13% reported having distressing memories about other unusually scary or difficult life experiences; 10% reported distressing dreams related to the war or other traumatic experiences. Fewer reported experiencing a recapitulation of traumatic events or feeling physically or emotionally distressed when reminded of such events.

TABLE 2.

PTSD Symptoms.

When in the military, ever in combat?	42.5%
Ever in some way involved with injured or   dying servicemen?	41.6%
Ever have distressing memories about the war?	14.2%
Ever have distressing memories about other   unusually scary or difficult experiences in your life?	13.3%
Have distressing dreams?	10.0%
Ever feel like it (the war or other scary and difficult   experience) is recurring or happening again?	3.3%
Feel physically or emotionally distressed or reactive   when something reminds you of it?	2.6%

Alcohol Use

Most participants were not currently drinking alcohol (see Table 3). However, about 17% reported drinking more than 5 days per week, and about 5% reported drinking more than five drinks per sitting. For some, health care professionals (12.2% of the sample) or others (5.5%) had suggested cutting down or have expressed concern.

TABLE 3.

Alcohol Use.

Item	Distribution	Percent
When was last drink of beer, wine,   whiskey, or other liquor?	< 1 week	39.2
	< 1 month	8.0
	> 1 month	52.8
How many days a week drink?	> 5 days	17.4
	3–4 days	5.7
	1–2 days	16.7
	None/less often	60.4
How many drinks in one sitting?	> 5 drinks	4.7
	3–4 drinks	8.0
	1–2 drinks	40.7
	None/fewer	46.7

Relationship Between Mental Health Symptoms

General anxiety symptoms were correlated with depressive symptoms (r = .47, p < .001). Endorsement of general anxiety and specific PTSD symptoms (having had distressing dreams, memories, recapitulation, and/or reactivity) were also correlated (r = .21; p < .001), but not depressive symptoms and PTSD. An ongoing grief reaction, (i.e., being “still affected” by a loss) was associated with more anxiety (t = 2.24, p < .05) and depression (t = 3.58, p <.0l).

Overall severity of alcohol use (quantity plus frequency) was not related to having a combat history (t = 1.48, ns), any PTSD symptoms (t = 1.29, ns), or ongoing grief reaction (t =.62, ns). Problematic drinking (in this analysis defined as three or more drinks at any sitting) was associated with the number of depressive symptoms (t = 2.04, p < .05).

Demographic Predictors of Mental Health Symptoms

Age was not associated with the number of depressive, grief, or anxiety symptoms, but it had a low positive correlation with the number of PTSD symptoms (r = .13, p < .05) and low negative correlation with problematic drinking (r = −.13, p < .05). Level of education was not associated with depression, grief, anxiety, or PTSD symptoms, but it was modestly correlated with problematic drinking (r = .26, p < .01). Individuals who lived alone versus with others were more likely to have problematic drinking t = 2.79, p < .01), as were individuals who were unmarried (t = 2.20, p < .05). We did not look at racial or ethnic differences due to the low numbers of non-Caucasians in our sample.

Acceptance Rates for Screening and Referrals

Of the 382 patients approached for interview, 316 completed the interview (29 refused and 37 were interrupted). Most participants (74.4%) were not rated by the interviewer as needing mental health services. Of those who did report symptoms warranting treatment, 60.7% (15.5% of the total sample) declined a referral for mental health treatment, and 39.3% (10.0% of the total sample) agreed to be contacted by a clinician.

An ANOVA with post hoc t-test with Bonferroni corrections was used to compare those not needing a referral, those rated as needing a referral who declined, and those rated as needing a referral who accepted. In general, those who were rated as needing a referral had higher levels of depression (F = 69.06, p < .001), anxiety (F = 37.93, p < .001), and grief (F = 12.41, p < .001) complaints, and those who accepted a referral had the highest levels of these symptoms (significantly different in post-hoc f-tests from those who were seen to need a referral but who refused). However, this was not the case for PTSD and alcohol use. Those who declined and those who accepted a referral had similar levels of PTSD (F = 3.70, ns) or alcohol use (F =1.84, ns).

DISCUSSION

Primary care clinicians have been encouraged to complete mental health screening as part of an annual primary care visit. Such screening is especially appropriate for older adults who may be less likely to recognize psychiatric symptoms or to ask for mental health referrals and more likely to present with somatic or nonspecific signs of depression. However, mental health screening is part of an increasingly large number of health, risk factors primary care clinicians attempt to include in increasingly short outpatient visits. As such, these screenings tend to be, of necessity, brief (e.g., 1–2 questions) and focus on a small number of issues, most often depression and alcohol use. For example, in our outpatient medical clinic, primary care clinicians ask about 10 risk factors (including diet, exercise, immunization, seat belt use, sunscreen use, advance directives, etc.). Depression is assessed by asking: During the past month have you been often bothered by: 1) feeling down, depressed, or hopeless; 2) little interest or pleasure in doing things?

In this paper we described our experiences with mental health screening performed in an outpatient Veterans Affairs medical clinic by a Master’s-prepared mental health professional, focusing on older adults not already in psychiatric treatment. The screening was brief yet more extensive than that typically included as part of primary care screening, focusing on symptoms of depression, anxiety, PTSD, grief, and alcohol use.

We found such screening quite feasible. Most patients were willing to be interviewed (all but 8%) while they waited to see their clinicians. The physicians and nurses in the clinic welcomed our participation. The patients did not report that the screening was too lengthy or intrusive. No patients reported adverse reactions to being interviewed. Of note, most patients did not report mental health symptomatology and appeared well-adjusted, at times even in the face of considerable medical illness.

However, a sizable minority did describe mental health concerns. Rates of depressed mood ranged from 9% (more than half the days in the past two weeks) to 18% (several days in the past two weeks) and are comparable to other studies. A small but not negligible number of. participants (~7%) reported feeling they would be better off dead or had thoughts of hurting themselves in the past two weeks. Given that older adults are at higher risk for suicide than other groups and that suicide is associated with depressive illness severity, it seems useful to ask both about affective symptoms and actual suicidal ideation (Raue et al., 2001).

Of interest is the fact that about one-third of the outpatients endorsed problems with insomnia and low energy, comparable to other studies (Ohayon, Zulley, Guilleminault, Smirne, & Priest, 2001). In our sample a minority of those with insomnia or low energy appeared to have disturbed mood or interest. It seems useful to screen for these specific symptoms as insomnia and low energy in and of themselves can significantly impair concentration, memory, and quality of life, and increase safety risks (Benca, 2001; Roth & Ancoli-Israel, 1999), yet can often be adequately treated with behavioral modification or short-acting pharmacotherapy agents (Ancoli-Israel, 2000). In our clinical experience, insomnia is one of the most common presenting complaints in our outpatient mental health clinic.

About one-fifth of our sample reported nervousness, worry, easy annoyance, or irritability at least several days per week, consistent with other reports (Arean & Alvidrez, 2001). As we did not complete full diagnostic interviews for specific anxiety disorders, we cannot comment on whether these symptoms represented clinically significant anxiety disturbance. However, it seems helpful to screen for anxiety symptoms as these may increase social isolation (Upadhyaya, Lyness, Cox, Seidlitz, & Caine, 2000) and physical symptomatology. In addition, in our clinical experience, some of these putative anxiety symptoms may, in fact, be part of a depressive syndrome, and in this study depressive and anxiety symptoms were moderately correlated. In older men, we find depression to present frequently with anxious/irritable mood and/or restlessness/agitation (as also described by Lenze et al, 2000).

Also of interest was the number (14%) of older veterans reporting distressing memories or dreams of traumatic experience. Such symptoms may be long-lasting, or may become more severe in late life in the absence of previously utilized coping strategies, such as work or alcohol use, and with the onset of medical illnesses (Schnurr & Friedman, 1997). In our clinical experience, these symptoms are amenable to treatment, even after many years. Psychopharmacology can aid with anxiety and insomnia secondary to PTSD. Psychotherapy can target strategies for managing distressing dreams and memories and for resolving existential issues raised by early trauma experiences. In this study, PTSD symptoms were not associated with depressive symptoms; specific screening for PTSD appears important for at-risk populations.

Similar to screening for PTSD symptoms in this veteran population with a high prevalence of lifetime experience of trauma, we found it useful to screen for grief reactions in an older adult population where nearly two-fifths had lost a close friend or family member in the past year and one-third were still affected by the loss; still being affected by the loss was associated with more anxiety and depressive symptoms. Most older adults adjust to such losses on their own, but such loss does predispose some to increased alcohol use, worsened health, depression, or even suicide, such that inquiring about recent loss is critical.

Most veterans in our study did not drink alcohol, consistent with other reports (Valenstein et al., 1998) and findings that older adults drink less than younger adults (SAMSA, 1998). A small number reported problem drinking, especially those who lived alone and were unmarried. While these adults were less likely to accept a referral, screening for alcohol use still seems worthwhile, given the potential risks of alcohol misuse and that some adults do stop drinking upon the recommendation of a health professional (Fleming et al., 1997).

In summary, mental health screening of older adults in primary care is important, although it can be challenging to assess depression in the context of multiple presenting problems (Rost, Nutting, Smith, Coyne, Cooper-Patrick, & Rubenstein, 2000). In this study, we found it quite feasible to screen for mental health problems and to work collaboratively with medical professionals, performing this screening outside of the general primary care visit while patients waited to see their clinicians. We found it helpful to screen for specific symptoms of depression including depressed mood, disinterest, insomnia, and suicidal ideation, and more “anxiety” related symptoms such as nervous or irritable mood. In this population, screening for PTSD symptoms and for grief reactions was revealing. Problematic drinking was less common, but could be easily screened with direct questions about frequency and quantity. It seemed useful to integrate mental health screening into the primary care clinic and to reduce the primary care clinician’s burden by providing such screening outside the primary care clinician’s session per se. Additional research to aid in targeting those most likely to benefit from screening, and to help clinicians in distinguishing true cases from false negatives or false positives (e.g., Klinkman, Coyne, Gallo, & Schwenk, 1998) is needed.

In this outreach screening program, only two-fifths of those with mental health concerns agreed to be contacted for mental health services. We did not assess whether treatment refusal related to lack of interest, lack of real or perceived need, or lack of experience with mental health services. In our study, medical patients with higher levels of depression, anxiety, or grief reactions were more likely to accept a referral, suggesting that higher levels of subjective distress played a role in accepting a referral. Yet, this was not the case for PTSD and alcohol use, suggesting that other factors may be related to treatment refusal for these patients. We did not have a diverse comparison group to consider whether age, education, race, gender, and veteran status may also predict treatment acceptance versus refusal. However, it is quite possible that our group comprised largely of modestly educated, male, combat veterans may be less likely to accept mental health referrals than other groups. In another study we found that persistent outreach efforts after initial screening can at times result in increased treatment acceptance (Moye, Rosansky, Llorente, & Jarvik, 2001) with such individuals.