Secondary Logo

Journal Logo

Review Article

Using technology to improve symptom management in ICU: a quality improvement study

Baker, Timothy MBBS, MRCP, FRCAa; Allan, Katie MBBS(hons), FRCA, FFICMb; Surman, Katy MB BChirc,; Navapurkar, Vilas MA, MBChB, DA FRCA, FFICMa

Author Information
International Journal of Surgery: Global Health: November 2020 - Volume 3 - Issue 6 - p e40
doi: 10.1097/GH9.0000000000000040
  • Open


This year international interest in the intensive care unit (ICU) has focused on how the COVID-19 pandemic has impacted on bed capacity, equipment, and provision of treatment options available. One area that has received less attention is the impact of an ICU admission on the individual. Patients and their relatives know all too well the consequence of their visit. One former patient published this description; “The ICU is a branch of hell and people that have been there know what I mean”1. This experience is quantified by multiple studies demonstrating long-term post-ICU morbidity with posttraumatic stress disorder, anxiety and depression developing in at least 20%, 30%, and 40% of patients, respectively2. We have yet to identify ways to improve or reduce the extent of this suffering2.

Despite the increasing age of ICU patients, and the increasingly significant comorbidities they arrive with, access to high quality ICU has allowed gratifying acute survival rates even with more severe presenting illnesses. Before the arrival of the COVID-19 pandemic, annual ICU admission rates were around 160,000 patients in the United Kingdom3 (a country with relatively low beds per 100,000 patients)4 and 5.7 million in the United States (with many more beds per patient)5. This has increased the number of patients surviving ICU and therefore also, the volume of physiological and psychological morbidity in the community. The surge in ICU admissions due to COVID-19, is likely to result in an increase in both the severity and mass of this burden.

Our understanding of the patients symptomatic experience in ICU is hindered because explicit recall of events is often compromised by illness and/or drugs. Patients report that their most distressing experience is the inability to communicate and that this means the suffering experienced whilst in ICU is often unrecognised or misinterpreted and hence untreated6. Patients describe these problems at follow-up7, meaning that symptoms experience are likely to be pervasive and that the implicit memories they encode may be contributing to the long-term morbidity we detect. Unfortunately, patient to clinician communication in ICU has been the focus of very limited research6.

We sort to better understand this problem, to find a way to improve our recognition of symptoms, and then to address them as they occur. We did this in an attempt to establish a mechanism that will allow us to determine if proactive symptom management results in a reduction in the incidence of morbidity seen at follow-up.


Sequential mixed methodologies were used in 3 development cycles to create the tool reported in this paper. In the first cycle, a focus group was established within which open questions stimulated discussion amongst former patients and relatives about how to improve communication in ICU. The discussion was observed and themes were identified by the lead author. These themes were used to develop a prototype software tool (myICUvoice).

The second cycle involved evaluating myICUvoice over a 9-month period in a subgroup of patients identified in cycle one as being most in need of improved communication—patients that were alert, orientated, and ventilated but physically unable to phonate due to the presence of a tracheostomy. These patients were randomized into an intervention (provided myICUvoice) or control arm (any standard communication tool). They were surveyed after discharge to evaluate and compare the perceived quality of communication between themselves and clinicians. The same survey evaluated the perception of all ICU patients admitted during the same time period but not meeting inclusion criteria for randomization. Staffs were also surveyed to ascertain their perception of the quality of communication with patients.

To develop and improve the software to be evaluated in cycle 3, a sample of patients were interviewed for constructive criticism after their discharge. Additional feedback was provided unprompted by staff, relatives, and patients. This feedback was collected and compared with the opinions of a wide range of expert clinicians (Table 1). A second iteration of myICUvoice was developed based on these findings.

Table 1 - Staff interviewed to refine software.
ICU doctors—including those newly qualified upto Consultants (the experienced senior doctors)
ICU nurses—including those newly qualified upto Matrons (the most experienced senior nurses)
ICU specialist nurses (responsible for long-term follow-up)
Health care assistants
Speech and language therapists
Psychological well-being practitioner
Allied symptom management specialists involved with ICU patients
Stroke physician and nurse consultants
ICU indicates intensive care unit.

Development cycle 3 evaluated myICUvoice when used outside of a study methodology. myICUvoice was available on an iPad in every bed-space in the ICU so that all patients admitted could use the software as a self-reporting symptom identification/tracking tool. Nurses were provided an education program to recognize that the concept of “voicelessness” described by patients in cycle 1 was apparent even in patients able to phonate. They were asked to ensure that every patient completed one symptom survey per day but patients were free to use it to report symptoms at any time. After 5 months both qualitative and quantitative methods were used to evaluate the clinical benefit of the tool. The qualitative study was conducted by an anthropologist with no prior experience of health care to minimize potential reporting bias. A grounded-theory methodological approach (including interviews and a survey) was used to gather the opinions and usage preferences of staff. Unprompted feedback continued to be provided by patients, relatives, and staff which was collated. The updated version of myICUvoice was also connected to a database. This meant we were able to quantify the number of users, the frequency with which the symptom survey tool was used and the incidence of different symptoms.

All 3 development cycles and their mix methodology evaluations were formally registered as a service evaluation project with the research and development department at Cambridge University Hospital.


Cycle 1 revealed that patients felt clinicians had limited understanding of the symptoms they experienced while in ICU and that optimizing communication would improve this. The group described a spectrum of “voicelessness”; some felt their complex conditions rendered expression of their symptoms invalid, others felt oxygen masks prohibited being heard, most recalled being too weak and/or tired to have the cerebral ability to find words or to repeat phrases when misheard, and others were orientated but ventilated with an artificial airway rendering speech physically impossible. All these components were exacerbated when communication was not in the patient’s native language. The patients suggested that routinely using a structured survey asking about symptoms would help prompt patients to better describe their problems. They explained that this was particularly important given the impairment of cognition they all experienced.

During cycle 2, 36 patients met the inclusion criteria. Eighteen were randomized to each arm. Fourteen remained in the control but 4 crossed over to the intervention arm because clinicians independent of the evaluation team felt it was in their best interest to use myICUvoice. Thirteen patients (36.1%) (7 intervention, 6 control) responded to the postal and telephone survey. The results, which have been presented previously, showed that patients meeting inclusion criteria (those who were unable to phonate due to ventilation) were more likely to be frightened and feel stressed than ICU patients outside this cohort and so may represent a group at high risk of post-ICU morbidity. The 7 patients randomized to have access to myICUvoice were less frightened or stressed and felt better understood than the 6 in the control group. Communication was felt to be effective in 57.8% versus 33% of the control group. It is notable that only 48.1% of all other patients (including all patients able to phonate) in the ICU felt that communication with staff was effective8.

Cycle 3 attempted to address the broader spectrum of voicelessness experienced by all ICU patients. After 5 months 226 patients had used myICUvoice (55.4% of the 408 patients admitted). A continuous stream of positive feedback was provided by patients and relatives9. Integrating the data generated by myICUvoice revealed inconsistent usage patterns. This prevents us from conducting a formal analysis correlating symptoms, specific conditions, and outcomes with sufficient rigor for publication. The early trends identified10,11 are, however, leading to the investigation of treatments not previously used. One such example is xerostomia, this symptom accounts for 30% of all reports and so now have synthetic salvia readily available for nurses to use.

The qualitative analysis of staff usage preferences, revealed that they used myICUvoice primarily with patients unable to phonate. myICUvoice was consistently being used for the patients that had met inclusion criteria in cycle 2. In this cohort staff described myICUvoice as incredibly beneficial12. The analysis showed that usability and content of the software needed to be improved before nurses would intuitively use it to address the broader spectrum of voicelessness. The qualitative researcher revealed why usage patterns varied; only 4% of nurses described an understanding of the patients perspective of voicelessness as a problem existing across a spectrum and was not isolated to those unable to talk. Nurses clearly understood that impaired communication was a function of the inability to phonate and that this needed to be addressed12. Despite training highlighting the feedback outlined by patients in cycle 1 which also included references to medical literature supporting the patients perspective; nurses still felt that their communication with patients able to verbally communicate did not need improvement.


The results presented and the COVID-19 pandemic meant that the authors have released the version of myICUvoice that was used in cycle 2. This version was designed specifically for the trachesomized patient and is available free of charge to be used by any other ICU/patient that requires it. It can be downloaded from the Apple App Store and is being translated into multiple languages to facilitate it being available internationally.

The decision to make it widely available was made after having demonstrated that for the physically voiceless patient, myICUvoice resulted in an improved perception of communication quality when compared to other tools available. Patients, relatives, and clinical staff emphatically expressed the desire to use myICUvoice and of its benefit12. Nursing staff continued to preferentially use myICUvoice unprompted in this patient group even when the tool was presented to them as a symptom self-reporting tool rather than a communication aid. Like many Alternative Augmented Communication (AAC) tools it has clear value for these patients; but we also acknowledge that communication techniques not dependent on AAC tools are used preferentially (such as lip reading) and more frequently by ICU clinicians13,14. This fact, and because we know with this approach we overestimate our ability to understand patients and underestimate the frequency of their symptoms15,16, is why more must be done if we are to improve outcomes for ICU patients. We will continue to develop the technology used in cycle 3 to meet our objective of ensuring that structured symptom evaluation becomes routine practice, not just the consequence of successful AAC tool deployment.

Although the software in cycle 3 failed to address the issue of voicelessness as a spectrum, it did allow the myICUvoice survey tool to identify, treat and compare symptoms experienced by 226 ICU patients. This provisional data has provided a better understanding of symptomatic experience of the ICU population; this is of enormous value as we have identified no other prospective analysis of this type.


The arrival of the COVID-19 pandemic stimulated the release of the myICUvoice. The impact of the pandemic on mental health is expected to be profound both at an individual and societal level. Fear and anxiety have been rife in the general population and this looks likely to be further compounded by continued global financial instability. In this context, patients admitted to the ICU will be at higher risk of long-term psychological morbidity. Anecdotal reports of COVID-19 survivors suggest that an increasing incidence of anxiety, depression, and posttraumatic stress disorder should be expected. This is further exacerbated by limited visitation rights to the ICU for family members; this worsens the impact of critical illness for both patients and their relatives17. Under normal circumstances help for these patients is available via social support networks, but this again is diminished due to social distancing measures and personal resources needed to support rehabilitation are also likely to be impacted. This new, long burning, constellation of problems needs serious consideration by acute care providers. Early interventions to improve patients’ experience on ICU are of pivotal importance in attempting to reduce long-term morbidity.

We acknowledge the limitation of using software in resource poor settings but here we would advocate the introduction of routine and repeated use of any symptom survey tool similar in structure to that used in this app. Any method to identify harmful symptoms and provide immediate treatment is likely to reduce the potential of long lasting morbidity that currently exists after discharge from ICU.

Globally we need to identify ways to increase awareness that voicelessness is multi-faceted and not simply the inability to phonate. We will continue to attempt to develop tools that are required to further study this area and to improve the current body of literature.

Ethical approval

The project was registered and approved as a service evaluation with the research and development department at Cambridge University Hospital.

Sources of funding

The Addenbrookes Charitable trust funded the development of the technology, provision of hardware, and qualitative researcher. The NIHR provided funding to support a research nurse development cycle 2.

Author contribution

Dr T.B. devised the main conceptual ideas, the project, supervised study implimentation and the proof outline. All authors discussed the results, reviewed the literature and contributed to the final manuscript.

Conflict of interest disclosures

Dr T.B. is the founder and CEO of Symptech Ltd, which retains the intellectual property rights for myICUvoice. The remaining authors declare that they have no financial conflict of interest with regard to the content of this report.

Research registration unique identifying number (UIN)





The authors acknowledge the following for their support. The technology deployed in cycle 2 was kindly developed (pro-bono) by Simon Burrows. The development of the technology now freely available was seed funded by the Addenbrookes Charitable Trust but would not have been possible without extensive additional (pro-bono) technical and logistical support from Mandrel Systems ltd and Positive Incline Ltd. Nadya Pohran conducted the qualitative research referenced in this paper but worked beyond her contract period to use this work to design and implement the improved user interface.


1. Alonso-Ovies Á, Heras LA, Calle G. ICU: a branch of hell? Intensive Care Med 2016;42:591–2.
2. Wade D, Als N, Bell V, et al. Providing psychological support to people in intensive care: development and feasibility study of a nurse-led intervention to prevent acute stress and long-term morbidity. BMJ Open 2018;8:e021083.
3. ICNARC. Summary atatistics. Available at: Accessed August 30, 2020.
4. Rhodes A, Ferdinande P, Flaatten H, et al. The variability of critical care bed numbers in Europe. Intensive Care Med 2012;38:1647–53.
5. Cox C, Carson S, Govert J, et al. An economic evaluation of prolonged mechanical ventilation. Crit Care Med 2007;35:1918–27.
6. Karlsen M, Ølnes M, Heyn L. Communication with patients in intensive care units: a scoping review. Nurs Crit Care 2019;24:115–31.
7. Jensen F, Thomsen T, Overgaard D, et al. Impact of follow-up consultations for ICU survivors on post-ICU syndrome: a systematic review and meta-analysis. Intensive Care Med 2015;41:763–75.
8. Baker T, Martin J, McCluskey R, et al. Communicating with mechanically ventilated patients: Can using technology help? J Intensive Care Soc 2017;17(suppl):93.
9. myICUvoice. What do our users think? myICUvoice—Impact. 2020. Available at: Accessed August 23, 2020.
10. Baker T, Surman K, Claridge MA, et al. Using technology to understand the symptoms associated with conditions presenting to critical care units and introducing pre-emptive strategies to treat them. J Intensive Care Soc 2019;20(suppl):178.
11. Baker T, Surman K, Claridge MA, et al. Early identification of psychological symptoms in critical care with a new software system helps target interventions aimed at reducing long term morbidity. J Intensive Care Soc 2019;20(suppl):137–8.
12. Pohran N, Neale W, Surman K, et al. An ethnographic assessment to evaluate how nursing staff responded to the introduction of a new software tool to improve the understanding of symptoms experienced by patients in ICU patients. J Intensive Care Soc 2019;20(suppl):179–80.
13. Albarran J. A review of communication with intubated patients and those with tracheostomies within an intensive care environment. Intensive Care Nurs 1991;7:178–86.
14. Hoom S, Elbers P, Girbes A, et al. Communicating with conscious mechanically ventilated critically ill patients: a systematic review. Critical Care 2016;20:333.
15. Randen I, Lerdal A, Bjork IT. Nurses’ perceptions of unpleasant symptoms and signs in ventilated and sedated patients. Nurs Crit Care 2013;18:176–86.
16. Wojnicki-Johansson G. Communication between nurse and patient during ventilator treatment: patient reports and RN evaluations. Intensive Crit Care Nurs 2001;17:29–40.
17. Newcombe V, Baker T, Burnstein R, et al. Clinical communication with families in the age of COVID-19: a challenge for critical care teams. Br Med J Opin 2020. Available at: Accessed August 17, 2020.

Intensive care; Critical care, Communication; Physician-patient relations; Professional patient relations; Mechanical ventilation; Endotracheal tube; Tracheostomy; Alternative and augmentative communication; Technology; Quality improvement; Self reporting

Copyright © 2020 The Authors. Published by Wolters Kluwer Health, Inc. on behalf of IJS Publishing Group Ltd.