Modeling return on investment for an electronic medical record system in Lilongwe, Malawi
Gerry Douglas
Dear Colleagues,
It’s a pleasure to have this opportunity to share some thoughts with the group about financial return on investment (ROI) for eHealth systems. I have been working on and off in health informatics in Malawi since spending a year working with the Malawi Ministry of Health as a volunteer in 1996. In 2000, motivated by my belief in the potential for electronic systems to improve healthcare delivery in Malawi, I founded a small NGO (Baobab Health Trust, http://baobabhealth.org) and started a pilot project to introduce touchscreen computers at the point-of-care in the pediatric department at Kamuzu Central Hospital (KCH), an 800-bed hospital in Malawi’s capital city of Lilongwe. Over the coming years this pilot work expanded into other clinical domains and other hospitals. The main thrust of Baobab’s work since 2006 has focused on building a point-of-care EMR system to support care and treatment for HIV patients, now managing roughly one third of all patient receiving antiretroviral therapy (ART) in Malawi. Other Baobab software modules include a radiology management system, specimen labeling system for laboratory specimens, Electronic medical record (EMR) for managing diabetes and hypertension patients, an admission / discharge system, and an antenatal care and maternity module. The patient registration system developed by Baobab in 2001 has issued roughly 1.6 million patient IDs across 29 hospitals. Baobab, now roughly 55 staff in total, is entirely Malawian operated and governed by a board of 6 Malawian trustees. I am now working in academia in the US, but continue to work with Baobab as a collaborator.
In 2009, recognizing the significant investment donor funding was providing to develop, deploy and maintain these system, I started thinking about sustainability. I had always framed my thinking about the systems Baobab was building as an intervention to improve patient outcomes, but never considered the implications on healthcare delivery costs. Working with Marco Cioffi, a consultant from Bain who volunteered six weeks of his vacation time (2009/2010) to help with analysis, we started to construct a simple financial model for hospital-wide use of EMR modules, initially focusing on a district hospital, and later on a central hospital. The model we constructed had two components, costs and savings, projected over 5 years. Using these estimates we constructed a simple net present value (NPV) model to determine if, overall, the costs outweighed the savings, or vice versa. Estimating the costs was relatively straightforward. We had been working at Kamuzu Central Hospital since 2001. Going department by department we estimated how many clinical workstation computers we would need, added in network electronics and servers, installation costs and training. Since we had been deploying system for almost eight years in Malawi we could estimate these costs relatively accurately. In addition to the initial investment costs we had to estimate annual costs to keep the system running. Based on the projected number of patients we were able to estimate the cost of consumables. Based on the power consumption of the equipment and the cost of electricity in Malawi we estimated running costs. Adding in some costs for maintenance and support we estimated the total annual running costs for the system. Now came the hard part. Our intuition was that the EMR would save the hospital money, but we needed to get beyond intuition and present a solid argument as to how and by how much. Marco combed the literature and was able to identify 12 district areas where EMRs had been shown to demonstrate cost savings. However, this was almost exclusively from EMR systems deployed in Western settings, and we were not comfortable with just carrying these figures over directly to the Malawi context. We were as conservative as possible, often cutting anticipated saving in half with the hope that this would represent a lower bound on the savings that could be realized at KCH. Finally plugging all this into a spreadsheet and doing a NPV analysis we were able to determine that not only did the savings outweigh the costs over 5 years, but the break-even point occurred in year 3. We were pretty happy at this point, but also aware that this was just a model based on lots of assumptions.
In 2011 I started collaborating with Julia Driessen from the graduate school of public health. We decided to refine the model built by Marco by bolstering the evidence behind the savings, update the cost component, and generally tighten-up the analysis. The main improvements were on the savings components. Looking at V1 of the model we observed that of the 12 areas of potential cost saving, the top four accounted for the vast majority of the savings. We chose to focus only on the top four, and ultimately reduced this to three areas, eliminating savings resulting from reducing adverse drug events (ADEs), as we had no information to support the prevalence of ADEs at KCH, or elsewhere in Malawi for that matter (currently studying this). We re-evaluated three areas of impact: length of stay, transcription time, and laboratory use. We collected data on expenditures in these categories under the paper-based (pre-EMR) system, and then estimated reductions in each category based on findings from EMR systems in the USA and backed by ambulatory data from low-income settings.
We compared these potential savings accrued over a period of five years with the costs of implementing the touchscreen point-of-care EMR system at KCH. Cost of procuring 151 touchscreen clinical workstations with associated thermal label printers and barcode scanners, network electronics, servers, installation and training was estimated at US $337,847 (this assumed no existing hardware infrastructure). Annual costs or electricity, consumable, maintenance and support were estimated at US $29,824 (higher in Year 3, as we budgeted to replace the deep-cycle batteries used for power backup). Estimated cost savings in length of stay, transcription time, and laboratory use totaled US$284,395 annually. When compared with the costs of installing and sustaining the EMR system, there is a net financial gain by the third year of operation. Over five years the estimated net benefit was US$613,681 (equivalent of the annual salaries of over 70 physicians). Full details here …
http://jamia.bmj.com/content/early/2012/11/08/amiajnl-2012-001242.full
If I were to put myself in the shoes of a hospital director in Malawi, presented with the possibility of installing an EMR system I would ask myself two questions … 1) Will the patients be better-off overall with an EMR?, and 2) will it be cheaper to deliver healthcare with an EMR? If the answer to both questions was “yes”, and this could be supported with some evidence, I would likely opt for an EMR. The question would shift from “Should I do it?”, to “How do I cover in initial investment costs?”.
In summary, I think modeling ROI is a valuable tool in articulating the value proposition of any information system. I look forward to feedback and some discussion.
Gerry Douglas.
Emanus Simeon
Hello Gerry,
Hope all is well.
Thank you for this insightful email.
I have seen a lot of NGO or Government (MHO) backed EMR implementations in the developing and 3rd world countries. These efforts often target the public healthcare systems. Is it possible to reasonably make a compelling argument to a private doctor in the developing work on why he/she should implement an EMR system? I will be interested in hearing back from any entrepreneurs who have been successfully development EMR business models.
Gerry Douglas.
--
You received this message because you are subscribed to the Google Groups "ict4chw" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ict4chw+u...@googlegroups.com.
To post to this group, send email to ict...@googlegroups.com.
Visit this group at http://groups.google.com/group/ict4chw?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.
Sarah Bird
This was really interesting.
Sincerely,
Sarah Bird
skype: birdsarah
email: sb...@alum.mit.edu
web: www.sarahbird.org | www.bonvaya.com
Mark Spohr
Mark
Gerry Douglas.
--
You received this message because you are subscribed to the Google Groups "ict4chw" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ict4chw+u...@googlegroups.com.
To post to this group, send email to ict...@googlegroups.com.
Visit this group at http://groups.google.com/group/ict4chw?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.
Nadi
Isaac Holeman
Hi Gerry,
This kind of cost modelling is so helpful, thank you for sharing! Although a 3-year payback is impressive, I think that between the academic decorum in your papers and your humility in fora like this email list, you don't convey some of the core reasons that I admire Baobab and think people on this list could learn so much from your work. You note that estimating savings was harder that accounting for your costs, but when I consider replicating your findings, it seems the more daunting challenge would be to deliver a quality EMR at the costs you describe. It's noteworthy that your labour costs are less than 1/4 as much as your hardware costs. Presumably this is because Baobab almost exclusively employs local Malawian staff and pays everyone local rates.
I'd like to comment on what I see as some distinctive circumstances and admirable yet difficult commitments that have contributed to Baobab's ability to deliver the quality at low cost. I'd welcome your feedback about whether my assessments are fair, and invite other practitioners on this list to think about how our organisations might be different if more of us made similar commitments.
Requirements for long-term ICT4D capacity building
- ICT4D organisation should be founded by tireless do it yourselfer who commits to a single country and specific projects until the job is done, without regard for lack of financial support in country or for said projects.
- In order to inspire the contributions of many staff and supporters, despite lack of funds, founder must be selfless and extremely resourceful. They may draw no salary for first decade. It will take this long for the organisation to be self-sufficient, given following constraints.
- All staff should be permanent residents of country of operation. In rare cases, hiring an ex-pat is permissible so long as they are paid standard local rates like everyone else. Locals who have worked abroad and come to expect higher international salaries will also be offered standard local salaries.
- Staff will initially lack many necessary skills or work habits, due to dearth of related work opportunities in country. Rather than recruiting experts from afar, founder will take time to just learn everything and patiently teach each employee until team consistently delivers high quality.
- Unwavering commitment to building capacity at ministry of health is necessary. In order for ties with ministry to not be eclipsed by wealthy yet short-sighted ICT4D projects, it is important that there be no other comparable organisations operating in country for first 5-10 years.
- Organisation must be distinct enough that staff are not poached too often by organisations with more foreign money. This will be easier if there are few if any comparable organisations in-country for first 5-10 years.
- Functional EMR should be in use for 5-10 years before cost study is undertaken.
Youngji Jo
ROI modeling for mHealth--some considerations
l Costing—Different activities and cost items across development phases (introduction > deployment > maintenance/ scaling up), which also often involve various organizations
l Saving—Savings from saved time, resources, and wage (compared to paper-based traditional systems)
eHealth (EMR)-inpatient setting vs. mHealth-community setting: For eHealth (EMR) effort, especially in the inpatient setting, higher patient volume can increase savings/revenues and therefore ROI. However, for mHealth which is community-based setting (in case w/o clear business model), benefits could be obtained mostly from operational “saving” rather than “revenue”.
However, are we really capturing all “saving” benefits? What we do and do not capture?
“Hard” benefits (ex. length of stay) vs. “Soft” benefits (ex. improvement in patient safety and quality from misinterpretation of handwriting): Here, we only focus on some tangible aspects of financial costs and financial benefits. Are there any systematic and quantifiable mechanisms capturing intangible benefits such as efficiency increases (ex. saved time in service delivery) ; spillovers ; enhanced capabilities (ex. improved daily productivity/ effective coverage of CHW) ; avoidance of wasteful or suboptimal strategies (ex. salary of data clerks) in estimating “savings” benefit of mHealth?
Scale of program vs. “break-even” points: Given considerable amount of initial fixed costs, small scale program, attempting to implement a health IT system often face relatively larger costs per patient and will require longer time periods to recoup revenue gains. Significant “saving” benefit could be obtained at a broader scale; It could be interesting to analyze the relationship between size of the program and the length of period taking “break-even” point. Similarly, this could be linked to identify change of '"marginal costs" over the course of scaling up.
TCO model à ROI model: Together with the current discussions on TCO for mHealth costing, the ROI modeling could add valuable aspects on “saving” benefits for mHealth cost-benefit studies.
Health Systems Program
International Health Department
Y...@jhsph.edu
--