It was early springtime in Australia when my son died. I took jasmine and dark-red sweet peas from my garden to his funeral and laid them carefully beside him, wondering how I could even keep breathing through the pain.

His name was Adam. He was 38, and more than six feet tall, but he was still my baby. His birth, as my first child, brought me to the most joyous life turn I’ve ever gone through; his death, the most shattering.

I’d spent the first weeks of his existence obsessing over him around the clock, preoccupied with the basics of survival and longing for a snatch of sleep. Now, in the first weeks after his death, I reeled through a twisted mirror image of the same experience.

It left me buckled over my kitchen sink, an awful, primal sound tearing from deep in my lungs. That sound – of keening – was one I’d heard just twice before: once from an animal, and then from a friend at her 12-year-old’s funeral. I hope to never hear it again.

The news of Adam’s death arrived by phone, in the middle of the night. The shock was so intense that when I tried to email people to let them know what had happened, my cold and shaking fingers made an eerie sound as they clattered on my keyboard in the dark.

In the weeks that followed, over many other nights, I’d come back to that keyboard, searching the internet for solace. I hoped that I might click on something, anything, that could help me get through my intolerable emotions, or make sense of my collapse.

But as I read through grief websites and other information meant for those in my position, I couldn’t help but sit in judgment. I’m a meta-scientist – my expertise is in assessing the strength of scientific evidence and writing patient information – and my professional self was always right beside me, peering at the screen.

Having that perspective made the process heavy going. The internet of grief is crammed with conflicting theories and advice. Even a quick scan of the scientific literature told me that, for several reasons, I was at very high risk of becoming the sort of person who makes others say, sadly, “She was never the same …”

But I could also see that my fate wasn’t sealed. I knew that people could find a way to carry infinite sadness and still have a joyful life. I couldn’t envisage that, but I set out to try, for my sake and my family’s.

So I groped my way, as best I could, through the competing claims of grief experts. There’s been a huge amount of research on bereavement – a search of just one biomedical database turns up about 10,000 papers published in the past 10 years – but I found it rife with unreliable results from tiny, problematic studies.

As is true in many areas of psychology and medicine, a mass of studies has formed into a giant smorgasbord from which one can pick and choose results to fit any narrative, even when the weight of the evidence points another way. Theories based on the flimsiest data – or on none at all – have shaped how people think about their own and others’ grief.

In those early days, I decided that I’d have to come back to this research later on, when I was in better shape, so I could work through the evidence more thoroughly. Some studies had already helped me make important decisions. Among the confusing swirl of data, I knew there would be research that could help me more, and others too.

The joy of my son’s birth had set me on the path to childbirth activism and epidemiology. It seemed fitting that I’d work on the science of grieving because of his death.

There are many things I want to understand. Some are quite specific: Do open caskets at funerals help more people than they traumatise? How is grief after sudden, unexpected deaths like Adam’s different from grieving after losses that come with weeks or months of warning? Other topics are broader: Might I have suffered more, or less, if I’d gone about my grieving differently, or would that have made no difference? What can communities do to reduce severe, prolonged grief? But I thought I’d start my journey with a look at how the path of grieving is described. For this article I’ve delved into the so-called stages of grief.

Denial, anger, bargaining, depression, acceptance. Before Adam died, I’d had a vague recollection that this whole idea had been debunked. But I quickly found that it was thriving. Indeed, when I turned to the internet of grief, I came across the stages everywhere, in versions that had sometimes been rejiggered or remixed.

The five-stage model wasn’t generated from data. It’s a theory, developed by the psychiatrist Elisabeth Kübler-Ross and published in 1969, that explains how people come to terms with their own impending death. I remember watching Kübler-Ross on television when I was a young teenager, absolutely enthralled, and later buying some of her books. By the time her second one was published, in 1974, Kübler-Ross had expanded her claims, such that the five stages would apply to the grieving process too. Families go through them once while their loved one faces death, she argued – and then they may again when that person has died.

Although I’d been a fan of Kübler-Ross’s work, mentions of this theory caused me stress when I was in extreme grief. Was my denial “stage” over yet? If not, how long did I have before I’d turn into a rage monster and scare my grandkids? Some grief websites warned that people can move backward and forward through the stages as they grieve – an idea that made me worry that the all-consuming despair could return. That possibility nipped at any sense of hope and encouragement I could muster: My anxiety would ebb, but then the internet whispered, It won’t last.

The five stages, so the theory goes, aren’t simply bidirectional either; they can also come on out of sequence or with stages skipped over.

“Keep in mind that these stages are meant to be descriptive and don’t necessarily apply to everyone or happen in the order presented,” Cleveland Clinic says. But other sources seem to argue just the opposite, suggesting that certain stages might be central to the grieving process: Depression may be overwhelming, psychcentral.com told me, but “this stage is a necessary part of your healing journey.”

And the time we spend in any stage, I learned from grief.com, “can last for minutes or hours as we flip in and out of one and then another.” If the five-stage model does not describe predictable steps along a typical path, then why even call them “stages” at all?

Some writers retrofit the substance of each stage to make them match up better to the reality of grieving. None of Kübler-Ross’s original stages, for example, seem to capture one of grief’s most common features: yearning for the person lost. But cramming new emotions and ideas willy-nilly into the model’s elements is nothing more than tinkering.

The last straw for me came at the website for Cake, an end-of-life start-up that boasts of being “founded by MIT and Harvard alumni” and has raised $7 million in venture funding.  At Cake, the “bargaining” stage has been rebranded as “bargaining and guilt,” in which you could be “feeling desperation, helplessness, and lose [sic] of hope.” The “depression” stage? Well, that could leave you “feeling overwhelming sadness, despair, and loneliness.”

Meanwhile, the University of Washington suggests that bargaining means “ruminating on the future or past, over-thinking and worrying, comparing self to others,” none of which has anything to do with bargaining, either as Kübler-Ross originally described it or as the word is generally defined.

I learned that my doubts were well founded. Plenty of researchers, practising psychologists, and expert panels have given up on Kübler-Ross’s theory; some have called for it to be “relegated to the realms of history.”

• The Atlantic report