A Cost Analysis of All‐Staff Emails in an Australian Hospital Network (CASH Study)

ABSTRACT

Background

Written communication has changed dramatically in the computer era. Email use has exploded in the last three decades, from the inter‐individual level through to mass dissemination. In the workplace, it has become the default mode of communication despite very limited critical appraisal. Specifically in the healthcare setting, its utility as a communication strategy for broadcasting information has barely been examined. The aim of this study was to quantify the use of all‐staff emails (ASE) across a public hospital and its umbrella Hospital and Health Service (HHS).

Method

An audit of one year's ASE was performed to determine the number, word count and readability using validated metrics. The median was used as a reference text. Seventy five randomly selected volunteers from the five occupation groups in the health service were timed whilst reading the reference text to enable an estimate of time and financial cost.

Results

Five hundred and six ASE were identified, with a median length of 622 words (95% CI 570 to 694). The median Flesch–Kincaid (FK) Grade score was 8.9 (95% CI 8.8 to 9.1). The calculated predicted salary costs if all staff read all ASE in 1 year were AU$6 116 386.47 (range AU$3 441 365.32 to AU$13 424 704.98) at the hospital level and AU$14 555 209.62 (range AU$8 258 984.03 to AU$31 778 109.75) at the HHS level.

Conclusions

A modest reduction in annual ASE burden could result in considerable savings for hospitals and hospital health services, which could be reinvested into patient care, procedures and staff wellbeing. Further research would improve the limited understanding of the impacts of ASE and build an evidence base for how to optimise their use.

1. Introduction

Language has been evolving as a communication technique for over 100 000 years. Written communication has advanced from symbols into writing systems and alphabets, whilst recording techniques have progressed from rocks to paper to electronic methods. Email was developed at The Massachusetts Institute of Technology in the late 1960s with the first‐ever email across a network sent in 1971 [1]. Uptake was very slow for decades, but email has now revolutionised communication. All‐staff emails (ASE) are a form of bulk, untargeted communication. In the corporate world, they are considered helpful in engaging employees, with proposed benefits such as improved workplace morale and mental health [2, 3].

Specific to healthcare environments, email is amongst the most common tools for written communication in 2024 [1]. Previous literature has highlighted potential benefits of email when compared with paging, paper lists, face‐to‐face or loudspeaker/overhead communication options. An observational study in the intensive care unit (ICU) environment suggested that the use of email for clinical communication enabled greater speed, reliability of communication and subsequently, a possible benefit to patient care and staff satisfaction [4]. Although email offers potential benefits, it also brings challenges for healthcare networks [1]. The importance of clear and concise communication in the healthcare setting has been increasingly recognised by training providers. For example, The Royal College of Surgeons of Edinburgh developed Non‐Technical Skills for Surgeons (NOTSS), an evidence‐based course designed to improve patient safety and surgical outcomes, which has been enthusiastically adopted worldwide [5]. The Royal Australasian College of Surgeons (RACS) lists communication as a core competency for the provision of high‐quality healthcare [6]. RACS recommends the use of ISBAR (Introduction, Situation, Background, Assessment and Recommendations), a widely used technique to encourage clearer communication in the healthcare setting [7, 8]. In contrast, it is unclear whether staff responsible for writing ASE in government‐funded healthcare networks are required to comply with such principles.

Necessity, brevity and readability of text are core concepts in written communication. Whilst the first two are subjective, readability can be assessed in a number of ways. Automated scores such as the Simple Measure Of Gobbledygook (SMOG) Score, Flesch–Kincaid (FK) reading ease score and FK reading grade are amongst the more commonly used to evaluate healthcare related information [9, 10, 11, 12, 13]. The FK‐reading ease score is rated on a scale of 0–100, where the model indicates it is easier to read texts with higher scores, whilst lower scores signify academic levels of text [12]. The SMOG and FK reading grades are estimations that approximate the year level the standard of writing is written towards, for example, a score of 12 would indicate that the text is written to a twelfth schooling grade level [12].

Whilst there have been some studies that assess the implications of emails on staff in healthcare networks, little is known about the length, readability and practical cost of reading these emails. To address this gap, we sought to assess the standard of communication in ASE and estimate the time and salary cost associated with reading these emails.

2. Methods

The study setting was an Australian tertiary hospital in the state of Queensland. A single consultant surgeon's email account was used as the data source. Every email received in the calendar year 2023 was preserved for analysis. The recipient email address (‘To:’) field was usually blank, presumably for privacy and/or security reasons, which prevented direct identification of an ASE. Hence, ASE were identified by manual review (Dr A.S.) of every email to ensure the title, salutation, content and originating email address were consistent with an ASE. A maximally conservative approach was taken. For example, emails were excluded if they were sent from an address that was unlikely to have sufficient authority/permissions to send to all staff, or if the title or content appeared targeted to a particular subset of hospital employees (e.g., clinical staff).

Queensland Health (QH) is the state government‐run provider and is comprised of the Department of Health and 16 Hospital and Health Services (HHS), each of which is made up of multiple public hospitals and other healthcare facilities. QH must annually report data about its workforce statistics in various formats, including the Minimum Obligatory Human Resource Information (MOHRI) data [14, 15]. The MOHRI is a whole‐of‐government methodology used for reporting and monitoring the workforce and is a count of actual employees that are active and paid in a position in the payroll system in a fortnight. A breakdown of the employees of varying occupations is available within this data. MOHRI data for both the test hospital and its umbrella HHS from 26 June 2022 was used.

Each identified ASE was copied as raw text and pasted into a readability calculator which follows the equations for FK reading ease score, FK reading grade and SMOG score [13]. Word counts were also recorded for all emails. Again, a conservative approach was taken, such that no text from attachments or hyperlinks were included. Where an email contained an embedded video as the principal communication strategy, the video transcript was used (where available) in place of email text. The normality of the distributions for readability scores and word counts across the emails was assessed via Shapiro–Wilk scores. Normality was not met, so the email with the median readability scores was chosen as the reference.

Consistent with MOHRI terminology, five occupation categories were sampled (medical officer [MO], administrative officer [AO], nursing resource group [NRG], health practitioner [HP e.g., allied health staff], and operational officer or other ward staff [OO]) [14, 15]. Statistician advice was sought, with a recommended sample of 15 people per category required to power the study adequately. After verbal consent, a total of 75 employees were timed whilst reading the reference email. The time taken to read the email (minutes, seconds) was recorded for all, and when known to the staff member, their wage band was recorded if they were willing to disclose it. Complete salary scales for each of these occupation groups are publicly available, and the 2023 rates were used for calculations.

To estimate the time and salary cost assuming that staff actually read all of the ASE, a time per word was calculated using the median and range of reading times across each occupation group for the reference email. The total word count for all emails was then multiplied by these values to calculate the median and range of total time for an individual to read a year of ASE. Taking a conservative approach, to convert this to a salary cost for each occupation category, we used the mode for hourly pay rate (which was always equal to or lower on the pay scale for both the median employee across the hospital or HHS as well as the median sampled employee for each occupation group). We then summed this using the MOHRI headcount. All costs are reported in Australian Dollars.

2.1. Patient and Public Involvement

Neither patients nor the public were involved in the conception, design or execution of this study.

3. Results

After screening, 506 ASE were identified. As described earlier, none of the readability assessment tools nor word‐count were found to have a normal distribution. The median FK reading ease score was 52.8 (95% CI 52.3 to 53.4) with a range of 32.1 to 81.9. The median FK‐Grade score was 8.9 (95% CI 8.8 to 9.1) with a range of 4.1 to 12. The median SMOG score was 12 (95% CI 12 to 12) with a range of 8.1 to 12. The median wordcount per email was 622 words (95% CI 570 to 694) with a range of 58 to 3148 words. The median readability score email was used as the reference; it had a wordcount of 1049. The total wordcount for the calendar year was 404 230 words.

Seventy‐five employees of varying occupation and pay ranks participated in a timed reading of the median readability scored email. The mode hourly wage for AOs was $41.91, $62.76 for HPs, $80.65 for MOs, $46.24 for NRGs and $32.36 for OOs [14, 15]. Table 1 shows the MOHRI headcounts for the hospital and the HHS as well as the mode salary for each occupational group.

TABLE 1.

Minimum obligatory human resource information (MOHRI) count of active employees in the payroll system in a fortnight for the Hospital and the Hospital and Health Service (HHS) [14, 15].

Occupation	Hospital MOHRI occupied headcount	HHS MOHRI occupied headcount	Mode salary ($/hour)
Administrative staff (AO)	1057.83	2858.93	41.91
Allied Health/Health practitioners (HP)	1093.3	2795.57	62.76
Medical officers (MO)	1097.10	2278.96	80.645
Nursing (NRG)	3322.58	7950.42	46.24
Operational staff (OO)	677.81	1455.69	32.36
Grand Totals	7248.62	17 339.57	—

Table 2 shows the estimated total time and cost of reading all the ASE sent in 2023 at both Hospital and HHS levels. When the fastest and slowest sampled reading times were multiplied across the annual ASE burden, the minimum reading time was 7.06 h, and the maximum was 141.29 h. Using the median reading time for each occupation group and multiplying by the mode salary for the respective occupation groups, an estimated total cost of reading the ASE for calendar year 2023 emails was $6 116 386.47 (range $3 441 365.32 to $13 424 704.98) at the hospital level and $14 555 209.62 (range $8 258 984.03 to $31 778 109.75) at the HHS level.

TABLE 2.

The estimated minimum, median and maximum reading speeds and costs across the hospital and the Hospital and Health Service (HHS) for each occupational group.

Occupation and mode salary ($/hour)		Reading speed (words/second)	Total annual reading time (hours/person)	Total cost ($)—hospital	Total cost ($)—HHS
AO
41.91	Min	0.12	13.74	$609 325.17	$1 646 784.46
	Median	0.18	20.10	$891 209.24	$2 408 614.65
	Max	0.46	51.38	$2 277 851.09	$6 156 203.58
HP
62.76	Min	0.07	7.51	$515 596.48	$1 318 381.11
	Median	0.14	15.53	$1 065 566.07	$2 724 654.29
	Max	0.25	27.83	$1 909 601.92	$4 882 855.43
MO
80.65	Min	0.06	7.06	$625 054.88	$1 298 400.39
	Median	0.11	15.09	$1 136 463.42	$2 360 727.98
	Max	0.19	21.64	$1 914 940.86	$3 977 826.65
NRG
46.24	Min	0.09	9.63	$1 480 083.90	$3 541 611.83
	Median	0.13	15.16	$2 318 798.11	$5 548 525.20
	Max	0.24	27.49	$4 223 172.72	$10 105 399.08
OO
32.36	Min	0.09	9.63	$211 304.89	$453 806.25
	Median	0.29	32.11	$704 349.63	$1 512 687.50
	Max	1.26	141.29	$3 099 138.38	$6 655 825.01
Annual estimated net total ($)
Min			47.59	$3 441 365.32	$8 258 984.03
Median			95.68	$6 116 386.47	$14 555 209.62
Max			269.64	$13 424 704.98	$31 778 109.75

Abbreviations: AO, administrative officer; HP, health practitioner; MO, medical officer; NRG, nursing resource group; OO, operational officer.

From a broader perspective, the estimated cost of all staff at the test hospital reading ASE for the year ranged from 0.0034% to 0.013% of the entire state's annual health budget of $101 billion. This proportion increased to 0.0082% to 0.031% of the health budget if all staff at the test HHS read ASE for the year [16]. Whilst we were unable to obtain state‐wide data on the breakdown of occupational groups, we note that the total full‐time equivalent positions across QH is 101 032.28 as of March 2023 [15, 16]. Assuming the test HHS has a similar cross‐section of staff compared with the remainder of QH, we estimate the maximum extrapolatable cost could be greater than $100 million if all QH staff read the ASE for the year [15, 16, 17].

Finally, to provide real‐world context, the estimated total (median) costs were converted to hypothetical individual staff that could be employed at mode salary, or procedures that could be done at average cost at the test hospital (Table 3). Staffing costs were calculated using modal hourly wages and an allocated 1976 h for a 1.0 FTE position.

TABLE 3.

Hypothetical additional staff or common surgical procedures that could be paid for by the estimated cost of reading the annual emails by staff at one hospital or one Hospital and Health District. Data was extrapolated across the HHS using costs of the respective surgical procedure at the test hospital [17].

Occupation	Number of positions (hospital level)	Number of positions (HHS level)	Procedure type [17]	Average cost of surgery	Number of procedures (hospital level)	Number of procedures (HHS level)
AO	73	175	Appendicectomy	$10 700	571	1360
HP	49	117	Cholecystectomy	$18 400	332	791
MO	38	91	Inguinal Hernia	$13 100	466	1111
NRG	66	159
OO	95	227

Abbreviations: AO, administrative officer; HP, health practitioner; MO, medical officer; NRG, nursing resource group; OO, operational officer.

4. Discussion

Excess delivery of information results in poorer communication [3, 5]. All‐staff emails presumably started with the best of intentions and surely flourished as a concept because they appear to be an affordable tool for mass dissemination of information in comparison with historical controls (e.g., printed flyers or letters). However, it is only the upfront costs, such as consumables or postage expenses, that are lower [18]. There are hidden costs associated with sending and reading emails [2, 3]. For example, in 2000 an audit of the time taken to access and read emails in an information‐technology company employing 120 people found the daily cost was GBP1,500, or GBP12.50 per person [18]. Emails are particularly susceptible to excess due to the lack of physical or financial constraints [19, 20]. There is no clear way to define ‘excess’ when considering raw numbers, length and recipient lists, but these metrics are easily measured and would be relatively easy to target.

From a clinician's perspective, any time spent reading emails is time not spent providing clinical care. It has previously been shown that a single ASE that took 1 min to read, when sent to a health service employing 10 000 staff, would equate to 4.5 person weeks of organisational time lost, regardless of any need for follow‐up actions [20]. Email burden can negatively impact patient outcomes via drowning of other priorities, especially where content is unrelated to patient care [21]. The scant literature available is mixed; increased burnout or mental health strain, both inside and outside of workplace environments [21] may be balanced by positive effects on morale and team building. The putative positive effects are not widely accepted and are difficult to quantify [20, 21].

Of course, some ASE will be necessary and can be a simple and effective way of mass dissemination of information. For example, using ASE to communicate ‘system downtime’ for critical infrastructure such as imaging or pathology systems helps clinicians to optimise workflow [4]. Such emails should be succinct, but in our sample, these ASE were amongst those that were written to the length and complexity of formal academic texts. It is foreseeable that when dealing with significant volumes of emails, staff may miss this important information that affects clinical care.

What practical steps can be taken to moderate the negative consequences of ASE? Firstly, if you can't measure it, you can't manage it [22]. As noted earlier, measurement of ‘excess’ ASEs could include raw numbers, length, appropriate targeting of recipients (i.e., if the information really needed to be delivered to all staff), and complexity [23]. Modest changes to each of these areas could considerably reduce the burden of ASE, bringing a number of advantages. These would include reduced opportunity cost (performing other duties instead of reading ASE), improved conditions for existing employees, actual cost savings that could be used to fund operations or staff positions and the resulting improvements in patient and clinician wellbeing from better staffing [23, 24].

Secondly, various non‐economic means to reduce ASE burden have been trialled across differing types of corporate structures with mixed levels of success [23, 24]. Limitations on who is authorised to send ASE, protocolising email lists for sub‐populations within workforces instead of all‐staff, and education around sending and managing emails have all been suggested in the Harvard Business Review [23]. Whilst these strategies may already be used to some extent across Australian health services, strict adherence to these principles should be demanded, as well as enforceable limits on word count, readability and number of attachments (including prior email chains) [24, 25].

Thirdly, existing clinical communication frameworks (e.g., ISBAR) may also provide strategies to reduce ASE burden. Such frameworks aim to teach the exchange of clear, succinct information for the relevant audience with the goal of improving outcomes in critical and high‐risk scenarios [6, 8, 25]. Again, extrapolating these simple principles to ASE would lend support to concepts such as tight restrictions on appropriate recipient lists, demonstrated expertise in evidence‐based communication methods by any ASE author, and a focus on minimising the cognitive burden of recipients. One practical example is the replacement of verbose subject lines with a simple title including an indication of whether action is necessary, urgent or that an email is purely informative [23]. This concise direction to readers would allow them to triage ASE in the context of other emails requiring attention in their inbox.

More controversially, a price signal for ASE, such as a cost per recipient or cost per word sent, would apply strong downward pressure on the ASE burden. Price signals have proven efficacy as a disincentive for negative behaviours such as smoking [26]. The median word count of a single ASE in our dataset was 620, which if printed on paper would require two pages (12 point font, 1.5 line spacing). For the test HHS with a headcount exceeding 17 000 people, this equates to over 34 000 pages of printing and 17 000 stamps and envelopes [14, 15]. We hypothesise that reinstating even a fraction of this cost would have a considerable impact on the volume and length of ASE. There are other constraints associated with earlier forms of written communication that appear to have been overcome but have actually just been hidden [23, 24]. One example is archiving; to an individual user in a large institution, there appears to be no limit to email storage, but there are considerable costs in the form of hardware requirements, electricity consumption and carbon footprint.

There are several limitations to our study. Firstly, our cost calculations assume that all staff actually read the ASE. Anecdotally, 100% of surveyed staff expressed a general dislike of ASE. We suspect an attempt to collect formal quantitative data may be unreliable, as some staff may feel uncomfortable reporting if they actually read ASE and if so, how much time they spend on it. However, ASE are sent by their author with the presumption that it is important and should be read by all staff, so we did too.

Another limitation is that we have presented a single‐centre snapshot of a single consultant surgeon's inbox. We note that hospital staff across the country and internationally may have highly variable experiences with their ASE burden. Future research, involving comparisons across facilities and from the viewpoint of other members of the workforce may yield insights regarding how generalisable the data from this study are. Similarly, there are considerable variations in some parts of the data that may be outliers for example, minimum–maximum reading times varying from 7 to 141 h. These variations were likely due to the modest number of participants, variable reading prowess, the Hawthorne effect and distractions in the physical environment when the test email was being read. Our study design could not account for variations associated with interruptions, and these distractions likely led to re‐reading, exaggerating the reading time. Conversely, this is reflective of daily life, but it limits the validity of the projected costs of ASE. Due to the constraints of our ethics exemption, we did not study specific groups of staff for example, Culturally and Linguistically Diverse (CALD) backgrounds, although it could be expected that these groups may have differences in reading time, degree of understanding and stress levels experienced.

By restricting the analysis to ASE, we are under‐estimating the potential time and cost burden of bulk emails overall. For example, the reference inbox belonged to a clinician, and we excluded 133 emails that were directed to all clinical staff (but not all staff generally). This pattern is likely to repeat through multiple other sub‐group levels. Of course, unnecessary email correspondence occurs down to the individual level. The low barrier to sending emails means it is easy to unnecessarily ‘reply all’, ‘forward’, ‘carbon copy’ or reply with an inane acknowledgement of receipt of a prior email, all of which take time to read and process and would lead to a dramatic underestimate of the actual penalty associated with emails [23]. This penalty should not be viewed in isolation; it is likely to have additional consequences to the individual (e.g., cognitive burden) and the institution (e.g., workflow, patient care) but this impact has not been quantified in the literature [20].

Finally, we have only considered the ‘demand’ side of the equation (we use the word as it pertains to market theory here, not actual demand). We do not have figures regarding the ‘supply’ side, such as the infrastructure or staffing costs necessary to write and send ASE, which would certainly raise the total cost associated with ASE. Whilst we did not have access to QH data, in 2015–2016 there were at least 4500 non‐clinical administrative positions in Australia's Federal Department of Health, with an average salary of $222 000 [27]. Anecdotally, in our dataset, many ASE were sent by persons in these types of positions. The salary cost savings from reducing the supply side of ASE could be comparable in magnitude to the cost of reading the ASE [27].

Future research should focus on strategies to reduce UNnEcessary ASE (UNEASE). The negative effects of UNEASE on the mental and physical health of staff, the care of patients, and aspects pertaining to climate change are all possible future study areas. An interventional trial using one or more of the recommendations above may be feasible.

5. Conclusion

ASE have been described in the literature as having both positive and negative effects in healthcare and non‐healthcare settings. There are limited data to quantify the costs associated with ASE. We demonstrated that with fairly conservative assumptions, the hidden costs of ASE are high and there is potential for large cost savings with even a modest reduction in ASE burden. Savings could be reallocated towards clinical endpoints and may also improve staff wellbeing. Additional research is needed to validate our findings and explore the ramifications of changing how ASE are integrated into hospital operations. Finally, this paper has a Flesch‐Kincaid Reading Ease score of 46.4, FK grade of 11.4 and SMOG score of 12. If you were a median medical officer at the test hospital, we estimate you have spent about 6.3 min reading our paper. Thank you for your attention!

Author Contributions

Arthur Samoylovich: investigation, writing – original draft, writing – review and editing, methodology, data curation and formal analysis. Adam J. Frankel: conceptualisation, methodology, investigation including timed readings, data curation, formal analysis, writing – review and editing.

Funding

The authors have nothing to report.

Disclosure

This work has not been completed with the aid of any research scholarship and is not linked to previous or external research. This work is an original article. All authors have completed the ICMJE uniform disclosure form at http://www.icmje.org/disclosure‐of‐interest/ and declare: no support from any organisation for the submitted work; A.F. received salary from Queensland Health for the last six years; A.S. received a salary from Queensland Health for the last year alone, and no other relationships or activities that could appear to have influenced the submitted work.

Ethics Statement

Research Ethics and Integrity of The University of Queensland (UQ) reviewed this study and it was deemed to be exempt from ethics review under the National Statement on Ethical Conduct in Human Research (Project ID 2024/HE001306).

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.