Cost analysis of COVID-19 test result notification using an automated messaging system compared to a staff caller practice in Alberta

Abstract

Setting

In Alberta, polymerase chain reaction (PCR) COVID-19 tests were an important step in detecting and isolating contagious individuals throughout the pandemic. Initially, a staff member provided results to all PCR COVID-19 test clients by phone. As the number of tests increased, new approaches were essential for timely result notification.

Intervention

An innovative automated IT system was introduced during the pandemic to reduce workloads and support timely result notification. At the time of the COVID-19 test booking and again following swabbing, clients had an option to consent to receive their test results via an automated text or voice message. Prior to implementation, a privacy impact assessment was approved, a pilot was undertaken, and changes to lab information systems were made.

Outcomes

Health administration data were used in a cost analysis to compare the unique costs associated with the novel automated IT practice (e.g., administration, integration, messages, staffing costs) and a hypothetical staff caller practice (e.g., administration, staffing costs) for negative test results. The costs of sharing 2,161,605 negative test results in 2021 were assessed. The automated IT practice demonstrated a cost savings of $6,272,495 over the staff caller practice. A follow-up analysis determined the cost savings threshold of 46,463 negative tests to break even.

Implications

Using an automated IT practice for consenting clients can be a cost-effective approach to reach clients in a timely manner during a pandemic or other instances warranting direct notification. This approach is being explored for test result notification of other communicable diseases in other contexts.

Introduction

Symptomatic testing, case investigation, contact tracing, and outbreak management were essential components of the COVID-19 response. In March 2020, widespread COVID-19 polymerase chain reaction (PCR) testing became available across Alberta to identify cases related to travel, symptoms, close contacts, or outbreaks. Testing criteria varied depending on the pandemic status. From March 4, 2020 to December 31, 2021, daily COVID-19 lab test rates ranged from 70 to 20,263 tests, on average there were 9302 tests daily (see Fig. 1 for the weekly test rates).

Fig. 1.

Daily count of COVID-19 laboratory test results reported in Alberta from January 1 to December 31, 2021

Prior to the COVID-19 pandemic, most respiratory disease tests were requested by a healthcare provider to guide individual patient management. The provider would typically collect a specimen, submit the sample to the lab, and await results. Once results were available, the provider would confirm the identity of the client, parent, or guardian; share the results; and outline further steps in-person or over the phone. During the COVID-19 pandemic, most testing in the community was done to guide public health measures related to isolation, contact tracing, and quarantining as well as outbreak management. With widespread testing, and Albertans legally required to isolate for 10 days or until they received a negative test notification, a new system for notifications was necessary.

Initially, an online screening tool was developed and used in conjunction with a phone booking system through Alberta Health Services (AHS) Health Link to identify those eligible for testing and to self-book appointments at COVID-19 swabbing centres under the authority of the medical officer of health. AHS staff were redeployed and hired to work in sites to collect samples, and lab capacity was increased to meet testing needs. For result notifications (i.e., positive, negative, or indeterminate), AHS staff contacted all tested individuals or the parent/guardian by phone with the results. As the pandemic continued, new approaches were essential to meet the increased testing and notification demands.

In March 2020, AHS Information Technology (IT) and Population and Public Health created an online booking tool for COVID-19 testing appointments (Alberta Health Services, 2022). This provided convenient and timely access to scheduling across the province. Subsequently, another innovation was developed to reduce the workload of the AHS results notification teams. This gave consenting clients their test results in a timely manner via an automated text (i.e., short message service (SMS)) or autodialer program. Informing clients of results in a timely manner was essential to containment efforts and supported patient- and family-centred care. Inter-organizational collaboration between AHS (the centralized provincial health services system), Government of Alberta (GoA), and provincial laboratories was essential to develop and implement innovations, and gain acceptance. Prior to the launch of the automated system, a privacy impact assessment (PIA: COVID-19 Online Self-Assessment and Test Booking Solution – Rapid Testing) was submitted to the Office of the Information and Privacy Commissioner of Alberta requesting permission to obtain client consent through the online booking tool to receive COVID-19 test results by SMS text or autodial message. To reduce the risk of breaching client confidentiality, messages were delivered to the phone number provided at the time of consent, and only the client’s first name was included, along with instructions for reporting incorrect message deliveries. An amendment to the Health Information Act was granted for COVID-19 test result notifications via automated messaging during the pandemic (Government of Alberta, 2021).

The aim of this project was to assess whether the addition of the automated IT practice for COVID-19 test result notification incurred a cost savings and whether it should be considered for other public health notifications. The first objective of this project was to assess the unique cost of providing COVID-19 test results using a hypothesized AHS staff caller practice versus the automated SMS/autodialer supported practice (i.e., automated IT practice). A retrospective cost analysis was conducted to evaluate the expenses of each practice from a publicly funded healthcare payer perspective. The target population for this economic analysis included AHS clients tested for COVID-19 over the time horizon of January 1 to December 31, 2021 (365 days). The second objective was to identify the cost savings threshold for providing COVID-19 test results during this timeframe. The findings will be used to support dialogue and decision-making on the appropriateness of automated notifications from a cost perspective.

Intervention

Automated IT practice

At the time of booking a COVID-19 test, all self-referring clients or their parents/guardians could consent to receive their results via autodialer or text message. Automated result notifications followed many of the steps of a typical infectious disease result–sharing process including confirming the identity of an individual, sharing results, and describing further actions required (see Fig. 2 for the automated messages). Clients consenting to SMS and autodialer with positive COVID-19 test results received their results and further instructions via an automated message, which was followed up with a call from a communicable disease control (CDC) case investigator to confirm their results, complete a case investigation, and provide anticipatory guidance. Clients with indeterminate test results were also called by a CDC case investigator to provide interpretation and instructions for additional testing. Clients with negative test results received an automated notification and no follow-up call (see Fig. 3 for the flow process). Notifications via SMS were sent out hourly. The SMS message delivery rate was set at 1000 messages per minute, 50,000 messages per hour, and 700,000–1,000,000 messages per day. The largest transmission was 7731 messages in approximately 8 min. Notifications via autodialer were conducted at a rate of 10 calls per minute and 70,000–120,000 calls per day. The delivery rates of both systems could be adjusted to accommodate operational needs.

Fig. 2.

Automated messages (SMS and autodialer) sent to consenting clients with negative test results

Fig. 3.

Process flow diagram of the automated IT practice and staff caller practice for COVID-19 test result notifications. Note: Within each stream, there were a small number of people who were unreachable. ind test, indeterminate test; ACeRT, Alberta COVID-19 Exposure Response Team; CD/OM, communicable disease and outbreak management; AD, autodialer

Staff caller practice

The staff caller practice was used as the comparator. AHS Health Link staff called clients with negative test results who did not or were not eligible to consent to automated IT practice, and CDC staff called clients with positive and indeterminate test results. Time to notify clients depended on the number of test notifications and staffing levels. The staff caller cost model was developed using human resources (HR) data and in consultation with Health Link. The HR cost item was the greatest expense of this model and was determined from median salary rates for the position and the maximum number of client notifications required per day for the month. Monthly staffing levels were formulated to provide notifications to clients within 24 hours and assumed the median rate of 1.5 calls to reach a client and fluctuated monthly according to COVID-19 testing rates. See Fig. 4 for maximum daily cases and the peak monthly daily maximum. For example, the staffing estimates accommodated 14,542 daily negative tests in May and 4602 daily negative tests in July. Additionally, this model followed a pragmatic and conservative estimate for training, vacation days, sick days, and staff attrition to determine the number of staff needed per month.

Fig. 4.

Amount of negative COVID-19 test notifications sent by automated IT practice and estimated staff required to complete this volume of notifications using the staff caller practice from January 1 to December 31, 2021. Note: blue bars, daily frequency of negative COVID-19 test results of clients consented to be notified by SMS or autodialer; grey bars, monthly blocks of maximum daily rates of negative COVID-19 test results clients consented to be notified by SMS or autodialer (Januarymax = 10,115, Februarymax = 7595, Marchmax = 9720, Aprilmax = 14,538, Maymax = 14,542, Junemax = 5094, Julymax = 4602, Augustmax = 6541, Septembermax = 11,937, Octobermax = 9954, Novembermax = 7401, Decembermax = 7774). Monthly blocks were calculated to estimate additional HealthLink staffing needed according to the number of additional clients with negative test results requiring calls

Cost estimate and model assumptions

This project estimated the costs associated with the two practices over a one-year period. Both the staff caller and automated IT practices required a CDC staff member to call clients with positive or indeterminate COVID-19 test results (see Fig. 3). Therefore, the costs associated with these cases were expected to be consistent. The processes for the clients with negative tests were unique in the two practices. For that reason, the scope of the cost analysis exclusively focused on the negative test result notifications.

Direct cost estimates of the automated IT practice and the estimated staff caller practice for the negative COVID-19 test of clients consenting to auto-messaging were calculated. A publicly funded healthcare system perspective was taken; therefore, only direct costs to AHS were included. Direct medical and non-medical costs to the patient, family, and productivity were excluded. Provincial health administrative data from multiple departments (i.e., finance, human resources, IT, Health Link, CDC) were used and reported in 2021 CDN$.

The automated IT practice cost model included actual costing data. The development costs included one-time startup costs (i.e., hub build, working group meeting and preparation time, software and licensing fees, and AHS IT analyst’s time to integrate the laboratory information system (LIS) with the Automated Contact Centre Hub system) and ongoing one-year implementation costs (i.e., IT analyst for maintenance, SMS, autodialer calls, HUB management, and Health Link calls of failed SMS/autodialer calls).

The AHS staff caller practice cost model included the hypothesized expected implementation costs associated with calling clients (i.e., clients consented to SMS or autodial with negative test results). The following assumptions were applied to the model:

• The costs associated with a positive or indeterminate test result notification were estimated to be equivalent across both practices; all positive and indeterminate results required a call from CDC for next steps.

• The costs associated with staff calling non-consenting clients were excluded, as these were consistent across both practices and out-of-scope.

• A flexible staff redeployment, starting employment, and ending employment process was assumed during the pandemic and may not be realistic (i.e., a one-month contract for caller staff).

Cost savings threshold

The cost savings threshold was assessed. The cumulative costs of each practice by number of negative tests were graphed and the point of intersection was identified as the cost savings threshold. For comparison and transferability of the results to other communicable disease testing, selected sexually transmitted and blood borne infections (STBBI) were used to estimate the potential cost of providing negative test notifications via SMS/autodialer compared to a phone call.

Sensitivity analysis

A follow-up sensitivity analysis assessed the degree of parameter cost variation of the automated IT practice that could be experienced, while maintaining cost savings. The sensitivity threshold was calculated by inflating the cost of the automated IT practice for a one-year time period to identify the tipping point where cost savings would no longer be observed relative to the staff caller practice. The sensitivity analysis provides additional information for decision-making on the generalizability and robustness of the cost analysis results.

Outcomes

In Alberta, 3,411,932 COVID-19 PCR tests were conducted on 1,794,799 people between January 1 and December 31, 2021, with a positivity rate of 8.08% (range: 0.68–35.09%). COVID-19 PCR test results included 2,899,807 negative, 275,776 positive, 236,349 indeterminate/other test. These numbers represent all those tested during this period, not only those eligible for the automated IT practice, consistent with the real-world implementation of this initiative. During the 2021 calendar year, approximately 76.3% (n = 2,603,412) of these clients opted to receive test results via automated message. Most clients consenting to automated messaging selected SMS (n = 2,491,461; 95.7%) over autodialer (n = 111,951; 4.3%; see Fig. 1 for the weekly test rates). Those who did not opt-in or were unable to receive an automated message (e.g., patient in hospital, visitors to many emergency departments, or in congregate living sites) received their results from a staff caller (23.7%; n = 808,520).

Automated IT practice

The cost estimate of the startup ($110,875) and implementation ($1,266,817) of the actual IT practice for negative COVID-19 test notification of consenting clients (n = 2,161,605) was $1,377,691 or $0.64 per negative test. See Table 1 for a breakdown of the items.

Table 1.

Items contributing to the $1,377,691 cost estimate of the actual automated IT practice for negative COVID-19 test notifications of consenting clients from January 1 to December 31, 2021

Phase	Item	Data source	Amount (CAD$)
Start up	HUB build*	Communication & Collaboration Technologies	$25,000
	Working group		$77,375
	Software and licensing fees		$0
	AHS analyst**		$8,499
Implementation	AHS analyst		$18,414
	Negative result SMS		$1,028,110
	Negative result autodialer call		$0
	Management overseeing hub		$27,960
	Health Link calls of SMS/autodialer failure***	Health Link	$192,333

All costs actualized in 2021 Canadian dollars

^*The Automated Contact Centre Hub build included the work completed to adapt the system to function with the AHS extract system, including pro-rated server fee

^**The cost for an AHS analyst to integrate the LIS with the Automated Contact Centre Hub system

^***Mean SMS/autodialer failure = 148 clients/day (2.1%)

Staff caller practice

The cost estimate of training and office equipment ($195,202), staffing ($6,158,111), and management, admin support, and office space ($1,296,873) were included. See Table 2 for a breakdown of items. The total cost estimate of the staff caller practice for negative COVID-19 test notification of consenting clients (n = 2,161,605) would have been $7,650,186 or $3.54 per negative test.

Table 2.

Items contributing to the $7,650,186 expected costs of an AHS staff caller practice for negative COVID-19 test notifications from January 1 to December 31, 2021

Phase	Item	Data source	Amount (CAD$)
Start up	Training for callers	Health Link	$60,137
	Additional computer and office supplies		$135,065
	Skype phone lines		$0
Implementation	Additional callers*		$6,158,111
	Additional management/administrative staff		$584,896
	Additional office space**	Finance	$711,977

All costs actualized in 2021 Canadian dollars

^*Additional callers were a combination of data coordinators (AS4) and administration support II (AS2) positions at a 1:8 ratio. Staff rate of pay was calculated as a POD including 1 AS4 and 8 AS2 positions (Johnston et al., 2022). Costs included salary, HR (OBP rate = 2% of salary) and benefit costs (20% of salary), and HR costs/costs to fill the position 7 days per week (1.68 FTE per 1 FTE)

^**Included rental, operating, and occupancy costs of the maximum number of FTE needed

Cost analysis

The automated IT practice for notifying clients of negative COVID-19 test results demonstrated a cost savings of $6,272,495 over the staff caller practice in 2021. The cost savings threshold for the automated IT practice was 46,463 tests (see Fig. 5). Assuming similar costs, result sharing practices, and client uptake (76%), conditions with ≥ 66,376 annual negative tests may observe a cost savings within the first year of implementation. The sensitivity threshold analysis identified a tipping point of 455%, in which the automated IT practice no longer offered a cost savings (see Fig. 6).

Fig. 5.

Staff caller practice and automated IT practice costs per negative test notification in 2021

Fig. 6.

Threshold analysis of the automated IT practice cost increase relative to the staff caller threshold

We explored the use of SMS/autodialer to inform clients of negative test results for other communicable diseases in order to examine the transferability of these results to other public health issues. According to the 2019 Alberta STBBI testing data, the cost of negative test notification via SMS/autodialer with 100% uptake was estimated to be $1,231,111 for gonorrhea, $1,193,346 for chlamydia, and $672,031 for syphilis. Currently, variation exists in negative test notifications according to testing location, the healthcare provider, and clinic protocol. The cost savings are less clear for this scenario, which lacks a standardized, centralized test results sharing system. However, an automated low-cost SMS/autodialer practice may facilitate the creation of such a system.

Implications

To our knowledge, this was the first study published examining the cost of automating COVID-19 test results to clients via SMS and autodialer. This study found a cost savings to the provincial healthcare system of $6,272,495 in 2021 by implementing an automated IT practice. The estimated cost per negative test notification case was $0.64 for the automated IT practice and $3.54 for the staff practice. The cost savings threshold was observed at 46,463 tests for the automated IT practice. This suggests implementation of automated IT practice for negative test notifications of conditions with ≥ 66,376 negative tests may observe cost savings within the first year (assuming 76% of clients are eligible or consent). Additionally, according to the sensitivity analysis, the cost of the automated IT practice could increase by 455% and maintain cost savings over the staff caller model, if ≥ 66,376 negative tests are expected.

Collaboration and partnerships

Intra-organization communication and collaboration within AHS was high during the development of the innovation with almost daily meetings among leadership and key employees (i.e., legal, IT, privacy, communications, CDC, medical officers of health, Health Link, labs). Regular, brief, and fast-paced working group meetings were critical, though they may seem costly. When considered within the context of developing a novel innovation over a short time period, the need for communication and collaboration was important. It is also a one-time cost and not required throughout implementation.

The development and launch of the automated test notification during the COVID-19 pandemic was necessary to provide timely result notifications. As in other jurisdictions, the pandemic expedited the development and launch to ensure the innovation was deployed as soon as possible (Mark et al., 2021). This acceleration also required inter-organizational collaboration between AHS, GoA, and provincial laboratories, which likely benefited from the streamlined centralized health system.

Client reach

The automated IT practice was used by most eligible Albertans. Most clients opting for the automated service requested SMS notifications (95.7%), although the autodialer (4.3%) provided an automated service to those without a smartphone or SMS. According to Statistics Canada, 88.1% of Canadians aged 15 years and older have a smartphone (Statistics Canada, 2018). SMS test notifications provided an additional communication channel, a more client-centred approach, a quicker way to receive results, and a process for clients to check messages at their convenience and save for future use (Bilello et al., 2019). Jong et al. (2021) reported Alberta clients with transient lifestyles and lower socio-economic status often rely on cellphones with text-only plans making SMS a better communication method. Additionally, clients expect healthcare services to use modern communication technologies (Jong et al., 2021). Some general client concerns associated with automated messaging include the loss of personal contact associated with manual calls and uncertainty of message privacy (Jong et al., 2021). With an additional communication channel of SMS/autodialer, clients may have a quicker, more convenient, and consistent way of receiving all test results.

Health service operator adoption

The obvious benefits of utilizing automated IT test notifications included the rapid notification of test results, especially in contexts with high volumes. SMS messages are considered a viable strategy for many health systems or clinics to provide timely results (Bilello et al., 2019). We found the automated IT practice to be more cost-effective than a staff caller practice for negative test notifications during 2021, if more than 46,463 test notifications were expected to be sent, depending on the rate of uptake. Additionally, providing clients with negative test notifications via an automated IT practice allows staff to focus their efforts on counseling clients with positive or indeterminate results. A reported concern of service operators includes the lack of certainty that clients have read and understood the message (Jong et al., 2021).

Other health interventions

Various health and wellness SMS interventions beyond test notifications have been implemented in Alberta. Most of these interventions include supportive messaging to subscribing clients to enhance health and wellness, improve mental health and addiction recovery, and reduce readmission to mental health units (e.g., Text4Mood (Agyapong et al., 2016), Text4Support (Shalaby et al., 2022), TextforHope, Text2Quit, Text4HealthyAging (Eboreime et al., 2022)). This innovation has demonstrated success and sparked the exploration of SMS and autodialer for other public health interventions, such as STBBI test and partner notifications.

Strengths and limitations

This study included several strengths and limitations. Strengths included the use of administrative data for the automated IT practice and estimates of the staff caller practice. Limitations included the conservative modeling of the staff caller practice with monthly fluctuation of staffing. Adding and removing staff monthly may not be realistic, and we assumed one-month contracts were possible via redeployment or casual employment. Additionally, costs of continuous redeployment and employment changes were not included. Other provinces have implemented other models that were not included in this study (e.g., “no news is good news” or request clients to check website portals for results). This study included the two AHS test notification models.

Conclusion

The findings from this cost analysis demonstrated a cost savings of $6,272,495 during one year of implementing an automated IT practice to notify clients of a negative COVID-19 test compared to the staff caller practice. The cost savings threshold was 46,463 negative tests and the sensitivity threshold found 455% was the cost savings tipping point. The results suggest that providing negative test results via SMS or autodialer for other communicable diseases may support sharing with clients in a timely manner. This also would be an important cost-saving method to consider for future pandemic events where immediate notification of results is required.

Implications for policy and practice

What are the innovations in this policy or program?

• This novel innovative intervention for automated test notifications provided:

  • - clients with their COVID-19 test results in a timely manner to support containment efforts and client- and family-centred care;
  • - a process for automated test notifications service to consenting Albertans while maintaining client confidentiality;
  • - a cost saving of over $6 million to the healthcare system within one year of implementation; and
  • - a case study/model to explore the use of automated test notifications for other public health conditions.

What are the burning research questions for this innovation?

• Additional research is required to:

  • - assess whether similar cost savings can be observed in decentralized healthcare systems;
  • - assess the potential for cost savings by offering consenting clients automated test notifications for other health conditions; and
  • - assess whether automated healthcare messaging, such as test notifications, can support person-centred care by aligning with client preferences, understanding the patient experience, and improving access and patient-reported outcomes.

Author contributions

All authors contributed to the conception and design. Material preparation, data collection, analysis, and writing of the first draft of the manuscript were performed by Loitz and Johnston. All authors reviewed, edited, commented, and approved the final version of the manuscript.

Data availability

COVID-19 testing statistics are available on the following Alberta Government website https://www.alberta.ca/stats/covid-19-alberta-statistics.htm#laboratory-testing. Only aggregated cost data used for this study are available (see Tables 1 and 2).

Code availability

N/A

Declarations

Ethics approval

N/A

Consent to participate

N/A

Consent for publication

N/A

Conflict of interest

The following authors were employed by Alberta Health Services: CL, JCJ, SJ, MD.