Article Text
PDF
Abstract
Background Glaucoma is the foremost cause of irreversible blindness, and more than 50% of cases remain undiagnosed. Our objective was to report the costs of a glaucoma detection programme operationalised through Philadelphia community centres.
Methods The analysis was performed using a healthcare system perspective in 2013 US dollars. Costs of examination and educational workshops were captured. Measures were total programme costs, cost/case of glaucoma detected and cost/case of any ocular disease detected (including glaucoma). Diagnoses are reported at the individual level (therefore representing a diagnosis made in one or both eyes). Staff time was captured during site visits to 15 of 43 sites and included time to deliver examinations and workshops, supervision, training and travel. Staff time was converted to costs by applying wage and fringe benefit costs from the US Bureau of Labor Statistics. Non-staff costs (equipment and mileage) were collected using study logs. Participants with previously diagnosed glaucoma were excluded.
Results 1649 participants were examined. Mean total per-participant examination time was 56 min (SD 4). Mean total examination cost/participant was $139. The cost/case of glaucoma newly identified (open-angle glaucoma, angle-closure glaucoma, glaucoma suspect, or primary angle closure) was $420 and cost/case for any ocular disease identified was $273.
Conclusion Glaucoma examinations delivered through this programme provided significant health benefit to hard-to-reach communities. On a per-person basis, examinations were fairly low cost, though opportunities exist to improve efficiency. Findings serve as an important benchmark for planning future community-based glaucoma examination programmes.
- glaucoma
- cost
- vision
- prevention
Statistics from Altmetric.com
Introduction
Glaucoma is the foremost cause of irrevocable blindness in the world. An estimated 2.7 million Americans aged 40 or older have some type of glaucoma, with a worldwide prevalence of approximately 64 million.1 Unfortunately, 50% of individuals with glaucoma remain undiagnosed because it is asymptomatic in the early stages and does not cause pain, and vision loss progresses slowly, initially affecting one eye.1–5 As a result, glaucoma is often unnoticed until the advanced stage.3 4 6 Early detection and treatment of glaucoma is important, since early treatment could potentially prevent vision loss, and glaucomatous damage in advanced stages of the disease is typically irreversible.
It has been estimated that 4.5 million individuals worldwide go blind as a result of open-angle glaucoma, and this number is expected to increase to 5.9 million by 2020.7 This figure is important because it underscores long-term morbidity and because blindness results in significant costs. In the USA, an estimated $623 is spent per patient per year when glaucoma is in its early stages, but considerably more — $2511 per patient per year — when glaucoma is advanced.8 Diagnosed glaucoma and optic nerve diseases average greater than $2000 per patient in annual medical costs.9 A 2013 report indicated that total glaucoma spending is $5.8 billion for the 2.8 million Americans who have been diagnosed with this disease.1 9
The US Preventive Services Task Force has concluded that there is insufficient evidence to recommend open-angle glaucoma screening in primary care settings.10 However, considering that the risk of glaucoma is higher among minority populations, targeting detection programmes to this high-risk group may be advantageous.3 11–19 However, there is little information on the cost of such programmes. A small number of published studies have evaluated the cost-effectiveness of treatment including detection, and several have measured the cost of detection alone (without considering effectiveness).4 20–25 However, the cost of a comprehensive community examination programme has not been established. We therefore report the detailed costs of the Philadelphia Glaucoma Detection and Treatment Project (PGDTP).
Methods
PGDTP was a community-based programme that targeted high-risk individuals between 1 January 2012 and 31 May 2013. Details regarding programme design, recruitment and inclusion criteria are published elsewhere, but in brief, PGDTP included free on-site educational workshops and ocular examinations at accessible urban locations in Philadelphia that serve predominantly older minorities (senior centres, churches, senior housing). PGDTP programme sites were located in neighbourhoods with high rates of poverty as compared with the national average because the low-income older minority population is at high risk for glaucoma and is typically underserved in eyecare.26 Inclusion was based on meeting one of the following criteria: (1) African–American ≥50 years, (2) ≥60 years of age and of any race or (3) family history of glaucoma.27 The educational workshops aimed to increase awareness about glaucoma and its symptoms, warning signs and risks. The ocular examination included a six-step on-site examination process (figure 1), culminating with an ophthalmologist assessment and the provision of on-site recommendations, referrals and follow-up appointments to individuals diagnosed with glaucoma or suspected of having glaucoma. Diagnoses captured included the following categories: open-angle glaucoma or angle-closure glaucoma (‘confirmed glaucoma’), glaucoma suspect, primary angle closure and other non-glaucoma ocular diagnoses. For each of these categories, diagnoses were recorded at the individual level and therefore represent a diagnosis made in one or both eyes. The prevalence of any glaucoma-related diagnosis (consisting of open-angle glaucoma, angle-closure glaucoma, glaucoma suspect or primary angle closure) was determined.
Examination process for non-dilated participants.
A cost analysis was performed using a healthcare system perspective in 2013 US dollars and included costs of the six-step examination and educational workshops. The analysis captured the value of staff time for training, travel, educational workshops and examination delivery, and non-staff costs for materials, supplies, mileage, van use and maintenance, and participant travel time and costs. Assumptions are shown in table 1. The cost analysis excludes existing self-reported cases of glaucoma (in other words, where a history of glaucoma diagnosis was stated by the participant during the medical history component of the examination).
Cost assumptions
Staff time costs
Examinations
The time spent for staff training, site visits to community partners and delivering educational workshops was obtained from logs kept by the project manager. Time spent travelling to workshops and examination sites was obtained from records kept by the project staff. Time required to deliver the examinations was collected by a trained research assistant who attended a subset of examination dates, at 15 of the 43 sites (36%). In total, the research assistant observed 1428 examination steps during the site visits.
Community health workers and ocular technicians were assigned to perform a majority of the examination tasks. Usually, one staff member was able to complete their assigned examination step without additional assistance. However, additional assistance was occasionally required. When this occurred, it was assumed that a project manager assisted and was present for half of the time required to complete the examination step. Rarely, a third staff member provided additional assistance with a step. This third staff member was also assumed to be a project manager, but was assumed to be present for only ¼ of the time of the first staff member. This approach to costing staff time was taken because the main purpose of the cost study was to understand which were the most time-consuming and expensive parts of the examination; thus costing in this manner was necessary to identify opportunities for increasing efficiency. Furthermore, costing based on staff members’ shifts would risk including research-related tasks, therefore resulting in overcosting of the examination.
The ophthalmologist assessment (step 5) included slit lamp examination (anterior chamber evaluation, indentation gonioscopy and funduscopy), measurement of intraocular pressure using the Goldmann applanation tonometer, interpretation of testing results, review of the diagnostic impression and establishment of a plan of management with each participant. When indicated, the ophthalmologist performed laser peripheral iridotomy or selective laser trabeculoplasty at the community site (laser costs not presented in this paper).
As a part of the programme, only those participants who had diabetes also received a dilated fundus examination (DFE) if they had not received one in the past year (based on self-report). Regardless of diabetic status, all participants had a fundus photograph taken of each eye. Completing the DFE increased the examination to eight steps since the test requires instilling eye drops then assessing the participant once he/she was dilated. We separately report examination time and costs for dilated versus non-dilated participants.
Other staff time costs
Other staff time costs included PGDTP general oversight/supervision, reminder telephone calls and travel time. For PGDTP oversight, it was assumed that the project manager was present at 10% of the educational workshops and 90% of the examinations. In terms of reminder calls, community health workers were responsible for completing and logging a brief telephone call for each participant a few days before his/her scheduled examination. Finally, staff travel time was collected from logs documenting the round trip time required to get to each site on the examination dates. The log specified which staff members rode in the equipment van versus those who drove separately. It was assumed that all trips occurred between the coordinating eye centre and the logged community sites.
Equations used to calculate the staff time costs of tasks are provided in online supplementary appendix 1. Wages for community health workers, medical assistants and translators were based on the 2013 US Bureau of Labor Statistics (BLS) wages for Philadelphia.28 The project manager wage was not easily ascertainable from BLS job categories; therefore, this wage was assumed to be 25% more than a medical assistant. Ophthalmologist wages were based on the National Institutes of Health (NIH) grant salary cap.29 A fringe benefit rate of 31.2% was applied to all wage costs consistent with the US BLS estimates at the time of the study.30
Non-staff costs
Non-staff costs included mileage costs of staff members who drove to various locations for workshops, trainings and examinations. Mileage was reported to the financial administrator at Wills Eye Hospital along with reimbursed amount. The 2013 Federal mileage reimbursement rate of $0.56/mile was applied to miles travelled.31
Additional non-human costs such as equipment, supplies, van maintenance and other expenses were also obtained from the financial administrator. Medical equipment was assumed to have 10 years of usability. We applied 20% of the equipment costs to the analysis, assuming 2 years of use and accounting for depreciation. The van was rented; therefore, depreciation was not applied to this particular cost.
Participant time and travel costs
Participant waiting time during the examinations was captured using a tracking form that was carried from step to step. Waiting time was calculated as total time per participant (end time minus start time) from the tracking sheet minus the staff time required to perform each step. Participant travel time to attend the examinations was captured during a brief survey administered at the conclusion of the examination visit. Travel costs include only transportation costs; participants travel time was not costed since the majority of the target population is non-working.
Base case total costs
Total costs of the programme were calculated as the sum of staff (time) and non-staff costs. These were referred to as the base case costs because the figures do not account for real-world variability. Base case total costs were then examined in relation to the number of new glaucoma cases detected to yield a cost per case identified. Since ocular disorders other than glaucoma were also found (eg, cataracts, retinopathy), we secondarily examined the total cost of the programme per case of any eye disease identified. Since a number of participants experienced suspect glaucoma (ie, examination findings were suggestive but not confirmatory of glaucoma), the cost per suspect diagnosis is separately reported.
Sensitivity analysis
To account for real-world variability, a univariate sensitivity analysis was performed to measure the effect of varying staff and non-staff cost inputs on total PGDTP costs. Staff time costs were tested by varying wages ±25% of the base case value, and also by testing the effect of adding up to 10 min of ‘downtime’ — time that personnel spent at the visits but were not performing any visit-related tasks — per visit. Additionally, van rental and office supplies costs were varied by ±25% of the base case value. The cost of medical equipment was varied from $0 (ie, donated to the examination provider) to the base case value. Similarly, the community partner fee (a modest payment to defray site-level administrative costs) was varied from $0 (ie, voluntary partners) to the base case value.
Results
Programme participation and diagnoses
A total of 1649 individuals participated in the project and received an examination. The sample has been described in detail elsewhere; briefly, 68% were female, 70% were African–American and 545 (33%) were newly identified as having a glaucoma-related diagnosis. The mean age was 69 years (SD 11).27
New glaucoma suspect diagnoses represented 330 (20.0%) of the examined cohort.
Newly identified and confirmed diagnoses consisted of 64 (3.9%) open-angle glaucoma and angle-closure glaucoma and 151 (9.2%) angle closure diagnoses (suspected primary angle closure and primary angle closure). There were 293 cases of non-glaucomatous disease detected: visually significant cataract (n=180; 10.9% of participants), diabetic retinopathy (n=39; 2.3% of participants), age-related macular degeneration (n=29; 1.8% of participants), other macular pathology (n=23; 1.4%) and other less common diagnoses (including ptosis, corneal pathology, suspicious nevus; n=22; 1.4%). The total number of ocular diagnoses detected among all participants, including glaucomatous and non-glaucomatous disease, was 838. Of note, participants could have received more than one diagnosis (ie, 838 diagnoses do not represent 838 participants).
Staff costs: examination time
Time and staff requirements for examinations with versus without pupil dilations are presented in table 2. On average, the intake and history step, the visual field test, and the ophthalmologist step each required 11–13 min.
Required staff time and costs for examinations*
The total cost for examinations without dilations was $64 680 ($40.45/participant). Costs associated with each of the initial examination steps without dilations ranged from $2.58 to $22.77, with the visual field test and the ophthalmologist step being the most costly at $5.00 and $22.77, respectively (table 2). For participants receiving dilation, there was an additional average cost of $6.31. The total examination cost for all participants (including both dilated and non-dilated) was $67 018 ($41.00/participant).
Additional staff costs: training, educational workshops, project supervision and staff travel time
Staff for this project comprised six medical assistants, four physicians, two community health workers and two project managers. Project staff attended 11 training sessions, 33 site visits and 60 workshops. Each activity lasted approximately 1 hour, and on average three staff members attended. Staff costs for travel, supervision, trainings, workshops and examinations are shown in table 3. The cost of supervising the programme (ie, by the project manager) was $17 081. Community health workers spent 25 hours making approximately 1510 calls to participants outside of initial examinations and workshops, which equates to $683 of personnel time.
Additional time costs: supervision, training, workshops and travel
Non-staff costs: mileage and equipment
Mileage reimbursement costs equated to $3697 ($2.24/participant), of which travel on the examination dates accounted for $1860, followed by travel to the educational workshops at $1067. Other non-staff costs such as van maintenance, supplies and medical equipment were added to the total examination costs (table 4). The total cost for equipment, supplies and other expenses was $109 035 ($66.12/participant).
Equipment costs and other expenses for the glaucoma detection programme
Total base case costs and per-participant costs for newly diagnosed ocular disease
Total base case programme costs including staff costs (trainings, site visits, workshops, examinations, supervision, telephone contacts and staff travel time) and non-staff costs (mileage, equipment and other expenses) was $229 082 across all participants. This translates to $139 for each participant examined. The cost per diagnosis is presented in table 5 and varies based on the diagnosis category. The cost per case identified of confirmed glaucoma (new cases only) was $3579, while the cost per case identified of a glaucoma suspect diagnosis was $694. Cost per case detected of angle-closure glaucoma diagnosis was $1517. When examining the effectiveness of this programme for any glaucoma-related disease detected through this programme, the cost per case was $420. The cost per any ocular disease diagnosed (total diagnoses, including glaucoma-related and non-glaucoma) was $273.
Cost per case newly identified, glaucomatous and non-glaucomatous eye disease
Participant time and travel costs
The mean total participant time for the examination itself, including waiting time and examination delivery by staff, was 92 min (SD 41). Wait time accounted for 36 min per examination, with the remaining 56 min representing actual examination time.
With respect to participant travel, 60% reported driving or using public transportation, 34% walked and the remainder used a taxi. Participants who used public transportation reported spending over an hour commuting to and from their examination. Those who drove had a mean total travel time of 40 min, but the mean cost of driving was $150 more than using public transit per participant.
Sensitivity analysis
A tornado diagram showing a series of univariate sensitivity analyses is presented in figure 2. Total programme costs were most impacted by varying wages. Specifically, the total programme cost ranged from $199 994 to $258 169, and adding staff downtime yielded a total programme cost range of $229 082–$273 689. Testing zero-cost community partner and medical equipment led to reduced total programme costs of $196 582 and $197 675, respectively. Varying the cost of the van and office supplies had only a modest impact on total programme costs.
Univariate sensitivity analysis tornado diagram.
Discussion
In this analysis, we found that the total cost of implementing PGDTP was $229 082 ($139 per participant examined). Our detection rates were comparable with prior screening programmes that focused on high-risk populations.27 However, a high number of narrow angle cases was detected. We attribute this to these participants coming from an underserved population with higher prevalence of cataracts, leading to narrowing of their angles. The cost per any case of glaucoma-related disease identified (newly identified open-angle glaucoma, newly identified angle-closure glaucoma, glaucoma suspect or primary angle closure) was $420. Also important to note is that our cost per diagnosis analysis excluded 100 participants who had a self-reported history of glaucoma, which was subsequently confirmed through the programme examination. Reconnecting these individuals to ophthalmology care represents an additional benefit of the programme.
Findings are important because detection and treatment can delay glaucoma progression and its associated costs (reported to range from $2162 to $2511 per year for late-stage disease).4 32 A recent study of US Medicare beneficiaries revealed greater healthcare service use among individuals with glaucoma and greater morbidity in individuals with glaucoma with visual disability; the latter cohort was associated with a significantly greater likelihood of depression and mobility difficulty.33 Prior studies have corroborated the relationship between vision impairment and disability, with one identifying a relationship between glaucoma and greater rate of motor vehicle accidents.34–37
With respect to the cost per participant receiving the examination, our figure of $139 is much larger than those reported in a 1990 UK study of community-based glaucoma examination which found a cost per participant examined of just £3.35 (equivalent to US$ 2014 costs of $11.72).38–40 The UK cost, however, did not include an ophthalmologist and did not include testing such as measurement of central corneal thickness and fundus photography. Moreover, the intraocular pressure was measured by a pneumotonometer rather than the gold standard Goldmann applanation tonometer, which was used in PGDTP. These differences could explain, at least in part, the difference in costs between the two programmes.
Our findings can be used to identify opportunities to improve PGDTP efficiency. Participants spent approximately an hour and a half at the examination, of which 36 min was waiting. This suggests an opportunity to streamline the examination programme. In particular, the check-in process, visual field test and the ophthalmologist assessments were time-consuming. The lengthy check-in was mainly attributed to technical issues with the computers, a long list of medications to be recorded or language barriers. These findings suggest that testing the computers during the planning phase, inputting medications prior to the examination date and having translators on hand (or a companion who can accompany their friend/family member to assist in translation) could hasten these steps. There were also instances when participants had to step away from the visual field test or ophthalmologist assessment because they were unable to sit on the stools for a prolonged period of time. Use of the Octopus 300 visual field machine (Haag-Streit, Bern, Switzerland), a shorter test, might mitigate this issue.
Staff credentials are also a consideration relating to cost. The community health workers, ocular technicians and ophthalmologists have relatively high wages. Engagement of medical assistants and optometrists may lower examination costs; however, the effectiveness of alternate staffing models in terms of glaucoma detection should also be considered. The Bristol Shared Care Glaucoma Study evaluated the ability of community optometrists to monitor patients with glaucoma and make valid measurements of visual parameters. It was reported that the community optometrists were able to provide quality assessments comparable with that of the ‘gold standard’ clinic references.41–43 Identifying optometrists who have experience diagnosing more complex ocular pathology would be important if this approach were used. Teleophthalmology is another approach that could reduce examination costs; the idea would be to conduct a simple teleophthalmology evaluation followed by more comprehensive glaucoma testing only if warranted.
Finally, if the programme were extended to include more participants over a longer duration of time, costs would decrease through economies of scale; that is, as the process becomes more refined and the fixed costs of the medical equipment are allocated over a larger number of participants and time, cost per participant would decrease.
The value of any community-based examination programme could be increased if it were targeted an at-risk population and its purpose was expanded to detect any eye disease rather than just glaucoma. Although detecting non-glaucomatous eye disease was not the main objective of PGDTP, there were many such diagnoses. As expected, the cost per case detected was lower when one includes both glaucomatous and non-glaucomatous disease ($273).
There are several limitations to this analysis. First, costs are underestimated due to exclusion of staff time to attend ancillary meetings, presentations, evaluation visits and video recordings. These are additional costs that are not related to conducting the cost study, but were excluded because the main purpose of this cost study was to understand cost drivers specific to the examination. Second, our analysis is limited to the initial examinations and workshops; treatments such as follow-up appointments for participants diagnosed with glaucoma and laser procedures were excluded. Third, we have not separated fixed costs (ie, costs of reusable equipment, such as the van and diagnostic devices) from variable costs (ie, costs for items or services that cannot be reused, such as personnel time costs and non-reusable medical supplies). Other limitations include a conservative estimate of ophthalmologist wage based on the NIH cap, the lack of a usual care comparison group and the short time horizon which precluded calculation of the programme’s long-term incremental cost-effectiveness. Finally, the sensitivity analysis focused on the impact of modifying examination costs, not population characteristics. Since our programme was targeted towards an at-risk minority population, the cost per participant figures presented herein are potentially less than what they would be if the programme was instead targeted towards a general population having a reduced glaucoma prevalence.
Conclusion
This is the first comprehensive cost analysis of a robust community-based glaucoma detection programme. Findings suggest that the health benefit of the programme is significant in terms of cases identified, and its costs are relatively small. Results from this programme serve as an important benchmark for planning and budgeting future glaucoma detection programmes. Opportunities exist to improve programme efficiency and decrease its costs.
Acknowledgments
The authors wish to acknowledge Katherine M Prioli for her analytical assistance with this project, and the Wills Eye Hospital glaucoma project team for facilitating the collection of cost data.
References
Footnotes
Contributors LTP served as the cost investigator for this study and oversaw all aspects of the cost study design, methods and analysis, as well as manuscript preparation, review and submission. MW served as the clinical investigator for the parent study and contributed clinical context to study design, methods development, and manuscript preparation and review. LH served as the research director for the parent study and contributed to study design, methods development, and manuscript preparation and review. HS served as the study coordinator for the parent study and contributed to manuscript preparation and review. PL served as a cost advisor for this analysis and contributed to manuscript review. JEC served as one of two expert collaborators at the US Centers for Disease Control and Prevention (CDC) and contributed to manuscript review. JBS served as one of two expert collaborators at the US Centers for Disease Control and Prevention (CDC) and contributed to manuscript review. DS served as the research assistant for the cost analysis and contributed to manuscript review. LJK was the primary clinical investigator of the parent study and contributed to study design, methods development and manuscript review.
Funding This work was funded by the US Centers for Disease Control and Prevention, grant #1U58DP004060-01. LJK reports grants from financial support: US Centers for Disease Control and Prevention (grant no 1U58DP004060).
Disclaimer The findings and conclusions in this paper are those of the authors and do not necessarily reflect the official position of the US Centers of Disease Control and Prevention.
Competing interests MW reports grants from US Centers for Disease Control and Prevention (CDC) during the conduct of the study; and reports grants from Iridex, Diopsys, Heidelberg Engineering, Allergan, OrCam, Merck, Partridge Foundation and Pennsylvania Department of Health, outside the submitted work. PL reports other funding from CDC outside of the submitted work. LJK reports grants from the CDC during the conduct of the study; and reports grants and personal fees from Allergan, from Aerie Pharmaceutical, Bausch & Lomb, InnFocus, Mati Therapeutics, and Diopsys, and personal fees from Alcon, Glaukos, Sucampo, Inotek, Sensimed AG, Alimera Sciences, ForSight Vision, Ocular Therapeutix, and Aerpio Therapeutics, outside the submitted work.
Patient consent No identifiable patient information is being presented, as all information is presented in aggregate only.
Ethics approval Wills Eye Hospital Institutional Review Board.
Provenance and peer review Not commissioned; externally peer reviewed.
Linked Articles
- At a glance