Title:Avoidant conversations about death by clinicians’ causes delays in reporting of neutropenic sepsis: grounded theory study.
Short Title: Reasons for delay in reporting of neutropenic sepsis
Authors: Catherine Oakley1,2, Cath Taylor2, Emma Ream3, Alison Metcalfe2.Addresses:
1Guys and St Thomas’ NHS Foundation Trust, Guy’s Hospital, Great Maze Pond, London, SE1 9RT, C Oakley Chemotherapy Nurse Consultant.
2Florence Nightingale Faculty of Nursing and Midwifery, King’s College, London, James Clerk Maxwell Building, SE1 8WA, A Metcalfe Professor of Health Care Research and Dean for research, C Taylor Senior Lecturer.
3University of Surrey, Faculty of Health and Medical Sciences, Duke of KentBuilding
Guildford, Surrey, UKGU2 7TE, E Ream Professor of Supportive Cancer Care and Director of Research
Correspondence to: , 0207 1884811 or 07989659857
ABSTRACT
Background: Evidence suggests patients delay reporting symptoms of neutropenic sepsis (NS) despite the risk to their life. This study aimed to elicit factors that contribute to delayed patient reporting of NS symptoms.
Methods: A constructivist grounded theory study used observations of chemotherapy consultations (13 hours) and 31 in-depth interviews to explore beliefs, experiences and behaviors related to NS. Participants included women with breast cancer, their carers (partners, family or friends) and clinicians. An explanation for patient delays was developed through theoretical sampling of participants to explore emerging areas of interest and through constant comparison of data and their coding. This entailed iterative and concurrent data collection and analysis. Data were collected until saturation.
Results: All patients who developed NS type symptoms delayed presenting to hospital (2.5 hours-8 days), sometimes repeatedly.Moderators of delay includedmetastatic disease, bereavement, fatalism, religious beliefs and quality of relationships with clinicians.There was an interplay of behaviours between clinicians, patients and carers where they subconsciously conspired to underplay the seriousness and possibility of NS occurring.
Conclusions: Findings haveimplications for health risk communication and development of holistic service models.
Key Words
Oncology, cancer chemotherapy, neutropenic sepsis, risk communication, relational care.
BACKGROUND
Mortality from sepsis (all causes) is 30%, rising to 50% for septic shock1. Each hour of intravenous antibiotic delay further increases mortality by 8% 2.Chemotherapy associated neutropenic sepsis (NS) causes approximately 700 deaths annually in the UK3. Studies from North America4and the UK5 suggest a mortality rate of 2-21% in patients who develop NS. A key explanatory factor is delayed help-seeking, despite access to dedicated 24-hour helplines 6-11. Preliminary evidence suggests patient risk factors for delay include having co-morbidities, advanced cancerand being 65 years6-7. Patient delays may also reflect nebulous symptoms associated with NS. Early symptoms are indistinct from common side effects of chemotherapy including diarrhoea or mucositis; fever is not always present3. More evident signs of NS, (e.g.hypotension, pyrexia or hypothermia, reduced urinary output, shortness of breath, confusion, unconsciousness, oedema) occur late in sepsis12-13.Late intervention at best delays treatment, and at worst increases the sepsis mortality risk 6,8.
Understanding the reasons for patient delay is fundamental to developing effective interventions to promote earlier presentation.We undertook interviews and observations with patients, carers and clinicians to explore the research question: “Which factors contribute to delayed patient reporting of NS symptoms? ”
METHODS
Context
Data were collected in the breast oncology clinic and chemotherapy day unit of a large UK cancer centrebetween November 2011 and February 2013. In this site, patients were informed verbally and in writing about recognising and reporting symptoms of NS prior to chemotherapy treatment (when doctors obtained consentand during nurse-led consultations). Patients were also given a card providing a 24-hour helpline for reporting symptoms, and information for other clinicians (e.g. General Practitioners/Accident and Emergency (A&E) staff).
Design
A longitudinal qualitative study involving interviews and observations with two samples of women, their carers and clinicians. Grounded theory is a methodology that informs both data collection and analysis with the aim of discovering theory to describe the phenomenon of interest 14: in this case the reasons for patient delayed reporting of NS symptoms. Constructivist grounded theory 15 is distinguished by reflexivity (to ensure clarity about the place of the researcher(s) as co-constructors of knowledge). Observation promoted closeness to conversations related to NS and in-depth interviewing was driven by continual reflection and increasingly focused inquiry to uncover unconscious, emotional or psychological reasons for behaviours. Ethical approval was secured from the National Research Ethics Service London - London Bridge (11/LO/0488).
Recruitment and Sampling
Grounded theory requires data are collected until no new information emerges (data saturation). Sampling is driven by findings from concurrent analysis (theoretical sampling). The sample initially comprised women whose chemotherapy had a 20% risk for NS16-17(prospective sample) and women admitted to hospital with NS (retrospective sample). Patients were recruited through clinicians; carers through patient participants; and clinicians through clinical leads. Written consent was obtained from all participants.
Women (prospective sample) were observed and interviewed multiple times throughout their chemotherapy treatment. This included observation of NS-related information delivery during initial consultations about chemotherapy, participation in informal interviews following these observations, and formal in-depth interviews typically after their first and their final chemotherapy cycle. The retrospective sample of women were interviewed following an episode of NS. Carers of the women (prospective and retrospective samples) and their clinicians also took part in observations and an in-depth interview. Interviews (informal/formal) explored experiences from participants’ perspectives of NS or witnessed events, e.g. the chemotherapy consultation or an episode of NS.
Data were collected by a chemotherapy nurse (CO) using an observation schedule and interview topic guide informed by literature about health care decision making and delay. Observations, informal interviews and reflexive thoughts 15were recorded in field-notes. Interviews were digitally recorded, transcribed and accuracy checked. Participants are anonymised and given pseudonyms throughout.
Data Analysis
Each interview and observation were coded, and codes developed into categories. Codes from new data were compared with existing codes and categories (‘constant comparison’)15, 18, in an iterative process of reading and re-reading transcripts, and recoding, before determining final categories and properties. This generated a complex, rich description, aligned with respondents’ experiences (grounded in the data). Hypotheses (generalised relationships among categories and properties)were developed and tested against prior and emerging data 14.A focus on negative cases (data not fitting the emerging theory) and behaviour variances further enhanced depth in analysis. The grounded theory was developed through application of a coding framework 19 (similar to Glaser’s theoretical coding families) 14 to weave the deconstructed story back together. The framework facilitated examining and explaining links between categories (see figure 1),to describe the causes, phenomena, contextual and intervening conditions, strategies and consequences associated with delayed presentation.Coding and analysis was led by CO, supported by all co-authors (all with qualitative expertise).
RESULTS
The study included thirty-five participants:13 women with breast cancer, nine carers, five doctors and eight nurses. Participants’ characteristics are summarized in table 1 and supplementary table 1. There were 31 in-depth interviews and thirteen hours of observations.
{Insert Table1}
Five women participated ‘prospectively’ in observations and in-depth interviews, and eight in retrospective interviews. Their median age was 56 years (range 31-74 years) and ethnicity was diverse. Five patients lacked an identified carer, nine held degree qualifications or above, and more had curable (n=9) than metastatic cancer (n=4). Three carers participated in observations and six in retrospective interviews. They included partners, adult children and friends; half lived with the patient. Carers (like patients) were highly educated. Clinicians included two consultant oncologists, two oncology registrars, four chemotherapy nurses, three breast nurses, one A&E doctor and one A&E nurse. Four clinicians participated in observation and in-depth interviews, four in the observation element only and five in interviews only. The majority of doctors were male, and nurses’ female.
All patients, except one,reported delaying reporting symptoms suggestive of NS - between 2.5 hours and “seven-eight days” - some on several occasions. Some women delayed reporting to the helpline, others bypassed it, eventually reporting symptoms to their GP or breast team. Three women never reported neutropenic symptoms.
Four interrelated categories were developed from the data; each containing 2-4 supporting properties (table 2). All categories related to the core category: ‘Playing down the seriousness and piling on the fear’.
{Insert Table 2 here}
Core category:Playing down the seriousness and piling on the fear
The core explanatory category conceptualizes the essence of patient delays. This encapsulates a dichotomy wherebyall participants colluded to underplay the seriousness of NS whilst clinicians heightened patients’ fears of dying (in attempts to encourage early contact). The influence of the core category is evident within each of the sub categories and their supporting properties.
Clinicians described delayed patient presentation with NS and subsequent harm as “frustrating” and: “… devastating...” (doctor 2) (Property 1.1: Holding in the frustration). Clinicians’also knewpatients feared “poisonous” “toxic” and “alien” chemotherapy. Melanie (observation) told doctor 1 she felt: “…routine panic and terror” and elaborated (subsequent interview): “…it has a dreadful reputation…”.Clinicians also realised patients felt backed into a corner:“…Patients think…as much as I don’t like being on chemo…I need to have it …”(CNS1).Melanie (observation)was: “…going to get this toxic cocktail, come what may” and Jackie (informal interview) knew without chemotherapy: “...there’s a possibility it will come back…decided to...endure it”. Clinicians’ consequently avoided fully explaining NS often to avoid frightening patients into refusing treatment (Property 1.2: Playing down the seriousness). Doctor 4 spoke of patients who were: “…terrified…you know…chemotherapy is…very beneficial…I … would say it’s important to call us rather than telling them that [the risk of death]”. Alice said: “…they minimise the side-effects….they don’t want people not to do chemo”. Patients also underplayed the risk. Hannah had: “... a gung-ho attitude...” to chemotherapy, which she: “...went into thinking that I wasn’t going to suffer like anybody else…”. Jackie was given the: “...worse scenario...I might not react...to the drugs as badly as... they say (light laugh)”.Sophia wondered: “…if it’s being a bit exaggerated” and Adanna believed: “...I’m not the first person to take it… (laughs)”. I saw some people having the chemotherapy...they were not laying on the chair...they walk by their legs, so it’s OK and I will be no different...”.
Clinicians described “piling on the fear” (Property 1.3: Piling on the fear)toconvince patients to act upon NS symptoms. Observed clinicians used emotive words like “it’s serious” “we worry” “you need to call us quickly” hinting at serious infections and the importance of reporting fevers. Words including NS, septicaemia, blood poisoning or blood infection were never used verbally or within written patient information. CNS2 said she told patients: “...in a very subtle way...it is the number one cause of death…” and CNS3 acknowledged the avoidance: “...without…telling patients…risk of dying from sepsis they already know”.Interactions did not address individual concerns or empower patients to manage side-effects. Checklists and dispassionate: “…stock phrases…” (doctor 4) were: “...automatic…” (CNS3) a: “…standard patter…” (doctor 1) (example: supplementary table 2). CNS1 told Sophia: “...you don’t have the ability to fight infection. It’s a life or death situation… a temperature of 38...”. CNS6 told Melanie: “…most important…is the temperature…infection is the one big thing”. Lynda described clinicians: “…drumming into me…a temperature…call it straight away…”. Patients felt processed. Sophia described: “…a spiel…a bit stupefying…” Paula lacked:“… time to think and ask questions…” and Jackie was on: “…a conveyor belt”. Clinicians also tried to counterbalance the harsher message by reinforcing beliefs about immunity. Doctor 1 (observation) advised Jackie due to her: “...general health...age…[chemotherapy side effects] will be very manageable”. Jackie (informal interview) said she was: “...fairly healthy ... I’m hoping my immune system is strong enough…”. Doctor 1 also advised Melanie (observation): “...I don’t think it will be as bad as you fear, it will be fine...”. Melanie (informal interview) felt neutropenic sepsis would be: “...very unlikely, I tend to be… robust…physical person ...”.
Clinicians stressed infections during chemotherapy could kill as a ‘scare tactic’. Doctor 1 explained: “…once they have…disobeyed… you really drum it home…people can die…”.CNS1was: “... firm and assertivewith Wendy...use the words life-threatening... you should have rung the hotline, this is nearly 24-hours... she won’t be doing that again...” (example: supplementary table 3). Alice’s friend said it is: “…difficult to stop piling on more fear…” Importantly, consenting for chemotherapy, visiting the chemotherapy day unit and hearing about NS heightened the reality of having cancer and thus fears of dying (Property 1.4: Reality hitting). Melanie felt: “…oh god…this is it now…”Jackie said: “…the fact…you have to ring somewhere just for that [infection]… I’d feel a bit awkward…it brings back home…being treated for something serious”.Sophia found the thought of reporting NS was: "...quite terrifying…” and made her realize she was: "…In a serious position…”.
Category two: Getting away with it
Doctor 4 interpreted NS as: “…low down on people’s list of things…they are worried about”andpatients were never observed to ask about it. Patients and carers were worried and most coped bythinking NS would not happen to them (Property 2.1: Ignoring the risk). Zeena’s niece knew: “…you could die from the chemo…you don’t want to accept it…”. Lynda said the: “...low white blood count hadn’t …concerned me...even though I knew it could happen ” and Adanna knew: “...you might die...don’t even think about it...”. Some avoided reading chemotherapy information. Hannah hoped: “…if I was unconscious…they…would find it chemotherapy alert card in my handbag” (examples: supplementary tables 4-6). Patients also often attempted to protect themselves (Property 2.2: Feeling protected from infection) through extreme precautions. Wendy kept: “… myself safe in the house…” and Jackie was: “…obviously washing my hands…”. Carers who feared and avoided thinking about death from infections were particularly protective in attempts to prevent these (example: supplementary table 7).
Patients initially ignored NS symptoms through fear (Property 2.3: Getting away with it). Doctor 5 said: “…if you don’t admit it, it’s not happening”.Hannah had: “…a dread of what it might be…you don’t want to face up to it”. Many patients tried to make NS symptoms go away by repeatedly taking their temperature (hoping this would resolve) or attempting to lowerthis. Doctor 3said an: “...ex-GP…was rigoring...took two paracetamol didn’t take…temperature”. Fayola:“...was… fine…Lemsip is working...”. Zeena tried: “…co-codamol…still the thing is there…”, Lynda was: “...trying to cool down…” and Wendy: “...ran a cold bath…”.
Prioritising normal life(Property 2.4: Normalising) was an avoidance tactic that contributed to delays.Jackie’s work: “…gave me some sort of normality”. Sam waited for her husband to take over childcare because she: “…wasn’t massively worried…”,Fayola: “couldn’t leave my work” andMelanie waited 24-hours and finished work before reporting a breast infection to her GP (examples: supplementary table 8 ). Delays may also be caused by a desire to be “normal”.Lynda: “... had friends staying...they all went to the pub... (laughs). Lynda’s Mother describedLynda’s concern “not to cause a fuss”.
Category three: Making the connection
NS was mainly explained prior to chemotherapy and was not on patients’ radar as important(Property 3.1: Not hitting my radar).Jackie was given the:“…worse…scenario…” and Fayola’s husband was: “…more worried about…sickness, the hair loss…we knew would happen…infection…probably might not happen…”. NS was a: “…difficult concept…to explain…” (Doctor 1). Clinicians (observations) warned about reduced immunity. Doctor 2 advised Adanna: “...white cells which protect against infection are low...” and CNS6 told Melanie: “...you’re more prone to infection...”.Patients heard:“…chemo batters your immune system…”(Helen)and:“…your white blood cells are…blasted…don’t have any resistance to infection…” (Alice’s friend); but were unaware of the riskiest timefor developing NS. Alice’s friend: “...didn’t take that in…”.
Patients often waited for a temperature of 38oC before calling.This behavior was reinforced because patients heard “feeling unwell” and “fever” as important simultaneous symptoms. Doctor 1 advised Sophia: “…feel unwell or…temperature you must call…” and CNS6 told Adanna to: “…take…temperature straight away if you feel unwell…”.Many patients assumed: “...infections…get... a dose of antibiotics…you are fine...”(Jackie).Fayola also: “...thought they were going to give me antibiotics to…use at home”and her husband did not: “…think it was going to be as serious…”.
Advanced symptoms of NS forced patients to take this seriously (Property 3.2: Getting a wakeup call). Zeena:“…was feeling so bad… shaking…legs killing me…couldn’t eat…Fever, when it was high I said…I need to call…”. By the time Fayola: “…finish work…and get to A&E it was something else entirely...I was really…sick.”.Theyrealized NS was not a simple infection. Paula said: “…it seems so minor the way they explain it…” andHelen reported: “…nowhere in this paperwork [patient information] was this illness”. Fayola knew: “…white blood cells…go down…you are prone to infection…” but NS was:“a wake-up call…my white cells was that low…I could have died…”. Terms like blood infection/poisoning or sepsis/septicaemia created better understanding. Some patients preferred explicit explanations of NS and the risk of death, whilst others preferred softer messages focused on personal benefits of acting quickly (examples: supplementary table 9).The word sepsis was enough to get Jackie to take it seriously: