On a summer morning, after several attempts to interview clients for an evaluation project, I arrived with a social worker at an overnight shelter. Finally, we had located Jules, who wanted to share her experiences with the program I was learning about. When we approached her and her friends, we noticed that she was sipping from a bottle of vodka.

Now this is certainly not an everyday occurrence in my life as an evaluator. Most of my days involve evaluation planning, liaising with stakeholders, and polishing up reports. The majority of my firm’s primary data collection is done by other members of my team, although occasionally I can’t keep my curiosity in check and I conduct a few interviews myself. For this project, I wanted to be closely connected to the program beneficiaries and ensure I had a very detailed understanding of their experience.

This project, to me, was a very big deal.

I had heard about it well before we were contracted to explore its impact on clients; I knew that it was incredibly innovative, that it was doing hard work that appeared vitally needed, and that I wanted to know more. It was without question that I wanted to claim the role of interviewer for myself. My academic background is in anthropology. While I knew that a life in academia was not in my future, I have been grateful for my training in social science methods. What I was doing in this project was close to participant ethnography and it rang true to me. Though short-term, as so many evaluation contracts are, I spent several days with a social worker trying to make meaningful contact with participants in this program.

The initiative we were evaluating was designed to support the most vulnerable residents of our city. Participants were invited to the program because they were found to be the most intensive users of publicly available supports. They all had very frequent interactions with police and the courts, relied regularly on overnight shelters and other housing supports, and were heavy users of emergency departments and emergency transport services. All had experienced a great deal of trauma in their lives and were living with addictions and mental health concerns, along with various severe physical ailments.

All of this background is leading up to a question of professional ethics — but a few more details first.

While we had relatively unprecedented access to various quantitative datasets – thanks to stakeholders determined to collaborate and establish the right consent and information sharing agreements – this evaluation would have been incomplete without investigating the lived experience of the clients. The quantitative data, compelling though it would prove, would not speak for itself. To really tell the story of this initiative, we needed to understand the clients’ journeys through services (or lack thereof) and how this initiative was different.

Jules’ social worker shared some of what she knew about her client. Jules had lived a frankly horrifying childhood in another province before moving to our city with her boyfriend; this boyfriend was also her abuser and her pimp. She was still young, in her early twenties, addicted to alcohol, spending nights in shelters, often beaten by people she considered friends, and yet still believing that she would be loved. Animal therapy was helpful for her, and she hoped to have children someday.

In the previous three years, Jules had been transported by ambulance 53 times. She had visited the emergency department 81 times; 11 of those visits were for head injuries. Her hospital admissions – times where she stayed overnight on an inpatient unit – were for reasons including pneumonia, a leg ulcer, pancreatitis and alcohol withdrawal. In one year, she had critical levels of ethanol in her blood eight times.

If it is not already clear, alcohol was an integral part of Jules’ life.

When planning our interviews, with the project team’s advice, we aimed to schedule times that would work best for clients who may be actively using intoxicants. For most, that ideal time was between 9 am and noon. Most of our interviews were scheduled at that time, although few were conducted smoothly. One client didn’t come home the night before and the on-site staff at his permanent support house couldn’t locate him. Another was able to share some of his experiences with me but became overwhelmed and we stopped the interview; his social worker spent more time with him after I left. Yet another client needed emergency care when we arrived at his apartment.

I should note that social workers were very involved in developing our consent and interview protocol. We did not invite any clients who had guardianship orders in place, or those who the social workers felt were not cognitively or emotionally capable of participating. Social workers discussed the evaluation and the interviews with clients before scheduling, and we reviewed information about the nature of the project, the risks and benefits of participating, before obtaining informed consent. Mental health support was always immediately available if it became necessary.

The “sober window” we were aiming for simply didn’t exist for Jules.

At this point in her life, she was drinking alcohol continuously, and substituting hand sanitizer or mouthwash when necessary. She regularly stole alcohol and products containing alcohol, leading to regular interactions with police and several warrants. At night, she was sleeping at a shelter; in the mornings, she would vomit repeatedly before starting to drink again. Without alcohol, she could not function.

So, when I met Jules at 9:30 on a summer morning, prepared with my recorder and questions, Jules was sipping vodka on a sidewalk surrounded by friends. We were both faced with decisions. Jules needed to decide if she wanted to speak with me, and where. I needed to decide if interviewing a client who was actively drinking alcohol was ethical.

I’ve shared this story with participants in project ethics training courses, and at an evaluation conference. People feel very strongly about their stance on whether I made the right decision. I was rapidly running through a mental decision tree.

Perhaps the most obvious question: can an intoxicated person give informed consent? In general, I would say no. If I was interviewing middle managers about their experiences of a workplace mentoring program and one of them was drinking or using other drugs, I would very likely reschedule. But what about when a person’s most functional state requires alcohol? Without alcohol, Jules was violently ill and could not function.

How could I be certain that Jules was participating voluntarily? She had previously told her social worker that she wanted to participate and was telling me that she still wanted to. But how confident could I be that she knew what she wanted? We were providing gift cards as honoraria – perhaps Jules was so desperate for that money that she would participate even though she didn’t really want to.

If I chose not to interview Jules, who would tell her story?

As an evaluator, I feel that one of our responsibilities is to give voice to people who may not have the opportunity or ability to share their stories.

You may have already guessed which path I chose. And you may be thinking that I am a clearly unethical evaluator and no researcher worth their degrees would be so cavalier. If that’s the case, I can’t blame you. Nowhere in my training did anyone suggest that I should pursue data collection with an intoxicated person. Although, to be fair, none of my early-2000s research training had really addressed this particular scenario. It had certainly provided me with examples of how researchers have exploited vulnerable people and done irreparable harm to them.

But I would like to defend my decision to proceed with my interview. I did consider several ethical points:

Voluntary participation: I truly believed that Jules wanted to talk to me. She had told her social worker so in previous days and repeated that desire when we met. At the end of the interview, she thanked me for speaking with her. The gift card, though helpful, was of a value less than what she could make through prostitution in the same time – I don’t think it was unduly coercive.

Informed consent: Again, Jules heard about the intention of our evaluation project and how the results would be used on more than one occasion. I do believe that though she was drinking, she did provide informed consent. I confirmed this more than once verbally before proceeding.

Respect for vulnerable persons: All the clients served in this initiative were highly vulnerable. By having Jules’ social worker at the interview, we ensured that any potential new disclosures could be followed up on, and that appropriate mental health supports could be immediately provided. This interview was one more opportunity for Jules to continue her relationship with her social worker, for whom she expressed gratitude many times during our conversation.

Giving voice: This point, of course, was what tipped the scales. If I chose not to interview Jules that morning, her voice and her journey would not have been included in our evaluation. As I’ve stated above, there simply was not a point during the day when Jules was not either violently ill or drinking; these were her two states of being at the time. To say that her voice was not valid, that it was less valuable than a sober voice, would have been unethical. You may disagree with me, but I stand firm that Jules’ voice needed to be amplified.

Had I not proceeded with the interview, I would not have heard her perspective on how much the program team had helped her. I would not have seen how dedicated her social worker was – our interview ended with Jules vomiting and her social worker calmly and compassionately cleaning up both her and the floor around her. I would not have heard about the severe trauma Jules experienced throughout her life. Understanding the deeply disturbing past that led to Jules, and other clients, being served by this initiative was fundamental to telling the story of its impact.

“I’m a hooker. I’m a drunk and a junkie.”

That’s how Jules described herself to me. She was physically and sexually abused for as long as she could remember. She feared seeking police support when beaten because her street friends who hurt her were who she considered her family. She did not want permanent supportive housing because she was terrified of being alone, even though when she was with others she was regularly beaten, robbed – even urinated on in the days before our interview. She drank until she blacked out, and often became violent during those blackouts. She desperately missed her incarcerated boyfriend and hoped that when he was released, he would love her the way she loved him. She enjoyed pet therapy and was effusive with praise and love for her workers.

Without that story, Jules’ collection of statistics looks frankly unimpressive. While her interactions with police services in the six months she had been receiving supports from this initiative decreased by 26%, her visits to the emergency department increased by 71% and her use of ambulance services more than doubled. Sharing Jules’ story puts context to those results.

I haven’t been faced with such a challenging decision since this project. But when it happens again, I am confident that I will rely on both my training and inherent respect for human dignity to guide my choice.