Understanding Why Patients May Not Report Suicidal Ideation at a Health Care Visit Prior to a Suicide Attempt: A Qualitative Study

In 2016, approximately 45,000 Americans died by suicide and a half-million individuals visited a U.S. emergency department for a nonfatal self-harm event (1). Outpatient medical settings provide opportunities for suicide prevention, given that nearly half of persons who die by suicide make health care visits in the month prior to death and the majority make a visit in the year prior to death (2,3). Therefore, the Joint Commission now recommends screening all patients in acute and nonacute care settings for suicidal ideation (4). Moreover, health care systems nationwide are implementing routine depression screening and monitoring (5), and depression questionnaires often include questions about suicidal ideation. Question 9 on the nine-item Patient Health Questionnaire (PHQ-9), for example, asks about the frequency of thoughts about self-harm in the past 2 weeks (i.e., “thoughts you would be better off dead, or of hurting yourself in some way”) (6).

However, the link between asking patients about suicidal ideation and the ability of health systems to help prevent suicide is not clear. For example, for detecting suicidal ideation (7), question 9 on the PHQ-9 performs similar to diagnostic interviews, and responses to question 9 are a strong predictor of suicide attempt and suicide death (8–10). Yet the sensitivity of this tool (among others) is only moderate. One study of patients who attempted suicide in the week after completion of the PHQ-9 found that one-fourth of suicide attempts were by patients who responded “not at all” to question 9 (9).

There are many reasons that patients may not report suicidal ideation to health care providers prior to a suicide attempt, but no research has qualitatively explored this topic from the patient perspective. The purpose of this study was to explore the perspectives of patients who made a suicide attempt after having reported no thoughts of self-harm during a health care visit prior to their attempt and to identify factors that may facilitate or preclude patients from disclosing suicidal ideation in health care settings.

Methods

Participants and Setting

Study participants were recruited from Kaiser Permanente in Washington State between July 2016 and October 2017 by using electronic health record (EHR) and claims data to identify patients with a probable suicide attempt within 60 days of having reported no thoughts of self-harm on question 9 of the PHQ-9. As previously described (11), probable suicide attempts were identified by a member of the research team during a manual EHR review to confirm participants had a documented self-injury or suicide attempt and were not currently hospitalized or institutionalized. A probable suicide attempt was defined by using ICD-10 codes for self-harm or possible self-harm. Eligible patients were mailed an invitation letter with the option to opt out by calling the research team. Otherwise the interviewer followed up by phone to schedule an in-person or phone interview about a week later. Patients were eligible to be interviewed until recruitment ended. Participants who agreed to be interviewed provided oral consent by phone or written consent in person. An initial $50 incentive was increased to $100 to increase participation. All study procedures were approved by the Kaiser Permanente Institutional Review Board.

Data Collection

Semistructured interviews were conducted by a psychologist who identified herself as a suicide prevention researcher. All interviews were audio-recorded digitally and securely transferred electronically to a professional transcription service. The interview guide was adapted from a prior study with the help of individuals with lived experience of suicidal behavior (12). The interview started by asking the participants to describe what happened at the time of their “injury” (replaced by “suicide attempt” once the respondent confirmed this term) and other open-ended questions about the circumstances—such as emotional state, relationships, work or school, and alcohol or drug use. Participants were then asked specifically about why they may not have reported having suicidal thoughts on the PHQ-9 prior to the suicide attempt and whether one of four specific descriptors applied to their situation at the time, including not experiencing suicidal thoughts, experiencing suicidal thoughts but question 9 did not match their experience, experiencing suicidal thoughts but did not want to tell anyone, or having experienced thoughts of suicide in the past but not while completing the PHQ-9. The interview concluded by asking patients what might have prevented the suicide attempt and specifically what the health care system could do differently. [The interview guide is available online as a supplement to this article.]

Data Analysis

Participant characteristics were summarized by using demographic information available in the EHR. Transcripts were reviewed for accuracy and uploaded to Atlas.ti, version 7. Data were analyzed by using a combination of directed (deductive) and conventional (inductive) content analysis (13). Directed content analysis was applied specifically to validate and further explore potential reasons why these patients reported having no thoughts about self-harm on question 9 of the PHQ-9 prior to their suicide attempt. Conventional content analysis was used to identify and describe emergent themes related to actions that a health care system could take to prevent suicide attempts.

After 10 interviews were completed, an investigator conducted an initial round of analysis to develop a codebook based on themes from the semistructured interview guide. Two team members (JR, UW) independently read all the transcripts to refine codes and choose quotations representative of themes. All investigators reviewed coded data to finalize themes and check conclusions against the data.

Results

Participant Characteristics

Forty-two individuals were identified as potentially eligible interview participants, and of those 26 (62%) agreed to be interviewed (10 refused, and six were unreachable). Most participants (N=22) were interviewed within 2 months of having attempted suicide (median=32.5 days, range 10–409 days). Of those interviewed, about half (N=15) were female and most were white (N=20), with age ranging from 18 to 63 (Table 1).

Reasons for Not Reporting Suicide Ideation

Many participants said they were not experiencing suicidal thoughts at the time they answered the PHQ-9. However, some explained that they had experienced suicidal thoughts around the same time. One participant reported, “There were those thoughts in general, but right when I was in the doctor’s office at that time, I was not experiencing those thoughts. I was just trying to get better.” Another participant explained, “Because it wasn’t really a super premeditated thing. Like, it was probably true when I filled that out that I was feeling that way [no suicidal thoughts], but it didn’t take very long to change that.”

In the case of suicide attempts described as completely unplanned, alcohol intoxication at the time of suicide attempt was identified as an emergent theme. About a third (N=10) of participants reported using alcohol at the time of the suicide attempt, and several described unplanned suicide attempts following an episode of heavy episodic drinking. “Well, it wasn’t anything premeditated,” said one participant. “I had been drinking heavily for quite a while and I drank quite heavily that day and that evening–a lot. Probably over [inaudible] with just the bourbon, and then I grabbed a handful of amitriptyline [tablets]. That’s all I remember really.” Another participant also believed alcohol played a role, explaining, “I'd been drinking and I got in an argument with someone and then I just—I don’t know. I just decided that I didn’t want to deal with any of it anymore so I took all the leftover prescriptions that I had.” A couple of participants also described a pattern of attempting suicide during periods of heavy episodic drinking. For example, one participant said, “I had been drinking again. This always happens when I drink, when I go on a binge.”

A few participants also explained that question 9 on the PHQ-9 did not match their emotional experience. One said, for example, “I guess I knew I wasn’t really in the mood to self-harm myself as much, but I knew I was depressed. But I didn’t want to say that because I didn’t know how else to put that. So I just kinda, like, left it out and kinda forgot about it.” Similarly, another participant described how he did not talk about his suicidal thoughts because he thought they would pass. “I’m not always honest about it, I guess,” he said. “The thought goes through my head but it just kinda passes by. It’s like, ‘Hey. Do it,’ and then I’m like, ‘No.’ Because it’s been like [that for] most of my life.”

Another participant, who reported prior suicide attempts, described his emotional state as other than suicidal prior to the attempt. “Like I said, for me it’s not—I don’t have the, days I’m feeling suicidal,” he said. If I feel suicidal, I’m going to act immediately. It’s not like I’m feeling suicidal, days prior. I don’t let it go that long. I’m just going to say if I’m feeling this way, I’m going to act on it. But if someone would say, ‘Do you feel that you’re maybe somehow emotionally dysregulated or emotionally out of control or emotionally overwhelmed or something like that?’ I would have said, ‘Oh, yes, absolutely.’”

Participants also described not reporting suicidal thoughts because of difficulty talking about them. For example, one participant said, “I guess just, like, I’m really bad with words it seems like, and when I get anxious, I get fumbled. So I just try to avoid talking as much, especially with stuff that’s gonna make me emotional and stuff.” It was also common for participants to describe fears related to the outcome of disclosing their suicidal thoughts, including concerns about confidentiality, anticipated stigma, overreaction, and loss of autonomy. One participant described feelings of embarrassment and stigma around disclosing suicidal thoughts: “It’s really embarrassing, you know what I mean? Especially in my world, where no one would expect me to do that. I just was so embarrassed. And I hate when they put it on record and it doesn’t look good and it looks like I’m crazy. I just didn’t wanna deal with that.”

Another participant described the fear of overreaction and potential loss of confidentiality, saying “I just didn’t want anybody to freak out. If they read that, who knows they won’t handcuff me and send off a whole bunch of sirens, you know what I mean? I would think the therapist would have to start telling other people, ‘Hey, this guy might hurt himself.’ I didn’t know if she’d call my wife or what she’d do.” Similarly, another participant described fear of repeating a past experience and loss of autonomy. “Everybody just freaks out and wants to get you hospitalize, and acts like you’re a danger, and that doesn’t really make me feel better. I don’t like that as soon as I say that that they want me to be monitored or so closely watched. I don’t want my privileges taken away or anything.”

A few participants described less common reasons for not disclosing suicidal thoughts to providers, including one who said she did not think her providers would help her, particularly because of her alcohol use. “I don’t feel they would have done anything,” she said. “I just feel like they would just be going, ‘Well, you’re an alcoholic and these feelings are normal and so on and so forth. There’s nothing we can do for you. You just have to get through it.’” Another participant mentioned wanting to make her providers feel as though they were helping. “It’s a guilt thing,” she said, “and I want them to feel like they’re helping me and the meds are helping even if they’re not.” Similarly, a few participants expressed a desire to avoid causing worry or stress for others.

What Health Systems Can Do

When asked what providers or health systems could do differently to help prevent suicide attempts, patients had a variety of responses, many about listening and expressing caring in a nonjudgmental way and without overreacting. For example, one participant emphasized the importance of listening and empathy: “I think listening to people fully before judging them is the most important thing. And trying to understand where they're at with things.” Another participant said, “If they showed interest in showing that they care . . . like, ‘Hey, you are going through some [crappy] times right now, is there anything that we could do?’” Similarly, another participant described wanting to be heard and given reason to hope, saying, “I just would like it if you could even call into an agency and just vent what’s going on, and see if there’s any alternatives to what else can I do besides this [feel like ending my life]? I mean, I don’t want to do this. I mean, I want to live. I may not have a lot to live for, but I don’t want to be buried or cremated.”

Another participant said that providers can encourage patients to be honest about suicidality by emphasizing privacy and offering help. “I guess just I would try to convince everybody who comes in there to be as honest as possible,” he said. “Emphasize you’re here to help them, you’re not gonna get them in trouble or anything like that, and tell them everything will be kept private unless you have their permission to tell other people.”

Several participants reiterated how overreaction may preclude disclosures about suicidal ideation. One said, “A lot of people don’t say anything because they’re afraid that if they say something that it’s going to cause a huge ruckus.” Similarly, another participant who expressed fear of the loss of autonomy reiterated how hospitalization is used too frequently in her experience. “I just always run into the issue where as soon as things start becoming difficult, they just immediately suggest that I go to the mental hospital,” she said. “I just cannot stress enough that it was not a good environment for me. And they still suggest that I go back, when it’ll just make things worse. It just seems like that’s one of their first options when it should be a last resort.” Additional recommendations for how health systems can help prevent suicide and some examples are summarized in Table 2.

Discussion

This research project identified factors that may preclude or facilitate the disclosure of suicidal ideation among patients in health care settings. The study was conducted by systematically identifying and interviewing patients who reported no thoughts about self-harm at a health care visit within two months of attempting suicide. As expected, we found that many patients were not experiencing suicidal ideation at the time they answered question 9 on the PHQ-9, which asks about the frequency of thoughts of self-harm in the prior two weeks. In some cases, participants reported answering the question based on how they were feeling at the time of their provider visit and may have not considered the two-week time frame of the questionnaire. Consistent with a prior report (14), a few participants also described how the wording of question 9 did not match their experience of feeling intermittently suicidal or emotionally out of control, highlighting a limitation of a measure designed to assess depression severity. Use of a validated suicide severity risk assessment that asks directly about suicidal thoughts and plans may aid identification of patients at high risk of suicide (15,16).

Other patients described experiencing suicidal ideation and chose not to report it to their provider. For these patients, fears surrounding the outcome of their disclosure about suicidal ideation were common, particularly the fear of being stigmatized or treated differently by their providers. Suicide-related stigma is common, and these fears are consistent with how suicide attempt survivors describe the decision-making process they use when considering disclosing suicidal thoughts to friends, family members (17), and health care providers (18). In this study, some patients did not believe disclosing suicidal ideation would be helpful and others avoided reporting suicidal thoughts to protect others, including their providers.

Participants offered suggestions for how health systems could help prevent suicide. Many described how listening and expressions of empathy are helpful, which is in line with prior qualitative research conducted among veterans (18) and with use of caring messaging as a suicide prevention strategy (19–21). Also consistent with prior qualitative findings (22), participants described how perceived overreactions, particularly loss of autonomy in the form of psychiatric hospitalization, are unhelpful and potentially devastating to patients’ faith in the health care system’s ability to provide patient-centered care for mental illness. This finding highlights a complex issue for providers, who may support hospitalization because of fear of the patient’s safety and their own culpability in the event of suicide (23). However, a core ethical challenge is that increasing the patient’s safety in the short run may not be the best long-term solution (24); the best protection for patients and providers may be collaborative development and documentation of a safety plan (23,25).

Heavy alcohol use at the time of suicide attempt was an emergent theme, particularly for suicide attempts that were described as unplanned. This theme offers a potential mechanism for identifying patients at increased risk of suicide, even when they report no current suicidal ideation. Substance use screening is becoming increasingly common, as more health systems implement the recommendation to integrate mental health into primary care (26). This may facilitate identification of patients at high risk of suicide alone or in combination with suicide prediction algorithms being developed to facilitate appropriate health care for high-risk patients (27,28).

The main limitations of this study were that participants were from one health care system and were not representative of all patients who report having no suicidal ideation before self-harm events. The purpose of this small qualitative study was to identify factors that may facilitate or preclude patients from disclosing suicidal ideation in health care settings prior to suicide attempt, but research among large, broadly representative populations is needed to generalize and extend these findings. In addition, some participants had difficulty recalling their experience filling out the PHQ-9 and needed reminding about how they answered. As a result, inaccurate recall may bias some results.

Conclusions

Many patients who attempted suicide following completion of the PHQ-9 did not report thoughts of self-harm on question 9, either because they were not experiencing suicidal ideation at the time or because of fear of suicide-related stigma, overreaction, and loss of autonomy. Screening for heavy episodic drinking may help identify individuals who make unplanned suicide attempts. Demonstrations of listening and expressions of empathy and caring may help patients overcome a fear of reporting suicidal ideation. This research could inform application of new approaches to address suicide prevention in health care settings, particularly as part of recommendations to identify patients experiencing suicidality and engage them in care (29).