Bahmni upgraded to v0.81 at SEARCH and Hemalkasa

SEARCH and Hemalkasa are chronologically the second and third implementation sites of Bahmni. Both are located in the Gadchiroli district of the state of Maharashtra in India. The region where these sites are located are remote, mainly covered with forest & hills, tribal and socio-economically undeveloped. It also comes under Red Corridor.

During June 2016, the ThoughtWorks team upgraded Bahmni at both the sites over 3 weeks. The timing was set to avoid the onset of monsoon during which the region often gets cut off and there are disruptions in electricity and internet.

This post is an update about the upgrade and how we went about planning and doing it at both — SEARCH and Hemalkasa.

SEARCH

Our development centre in SEARCH during the 2 weeks

A brief about SEARCH

SEARCH (Society for Education, Action and Research in Community Health) is an NGO founded in 1985 by a doctor couple, Abhay Bang and Rani Bang. It is inspired by the life and philosophy of Mahatma Gandhi and its mission includes community health care, research and training. They started the hospital in 1993. The hospital includes

  • Daily OPD with 200 patients/day
  • 20-bed IPD
  • Periodic surgery camps with surgeries performed at on-premise Operating Theatres by visiting volunteer surgeons
  • Monthly mental health clinic and daily dental clinic
  • In-house Laboratory, Digital Radiology, Ultrasound and ECG.

Previous Deployment

Bahmni was first deployed at SEARCH in June 2014 and was rolled out at Registration, Lab and Billing. That was the only deployment done at SEARCH before this upgrade.

A nice slide deck by Gurpreet about the lessons learnt from the deployment can be found here.

So, when we went there, we saw a really old Bahmni. It was so old that someone would mistook it for another product! ;-)

Reason for Upgrade

A couple of members from SEARCH visited JSS last year to see how the clinical part of Bahmni getting used AT JSS. After some deliberation SEARCH decided to upgrade and rollout the Clinical app for the doctors and nurses in OPD and possibly IPD.

Key deliverables planned with this upgrade were Clinical rollout and PACS integration.

Preparation & Planning

A BA and Implementation Lead from Bahmni team started analysing the requirements, the ones that had come up in previous deployment as well as the new ones. A test environment with the anonymised database dump and Bahmni product version from production was setup. Based on analysis and next level of discussions with the SEARCH team, a few minor modifications to the product were planned. Based on features and timeline, Release 0.81 was decided to be deployed.

As soon as the release candidate build was available, the implementation team started to identify steps needed for the upgrade and the new configurations. Multiple showcases were done to the SEARCH team remotely and based on feedback received, improvements were made.

As this was an upgrade after a long time and included clinical rollout, a team of a BA, a dev, an implementation Lead and an implementation engineer was planned to travel for two weeks.

A schedule was prepared for different activities to be done during the 2 weeks.

On-the-site / In-the-field

Reaching SEARCH from Bangalore took a couple of hours by flight followed by a 3 hour road journey. We met Nikhil there, our point of contact and EMR coordinator at SEARCH.

User training

Followed by a simple and delicious lunch, 4 of us got divided into two pairs. One pair focussed on setting up the training laptop and the other pair worked on coordinating with Nikhil to check readiness in terms of hardware, planning and availability of users for training.

Based on work timings and availability of the users, Nikhil planned the user training.

During each training session, we first explained the features by demonstrating them on a big screen using a projector which was followed by hands-on practice by the users.

Since nurses had never used computers before, Nikhil used an online game to train them to use the mouse. For keyboard and data-entry training he used google forms. The nurses seemed very enthusiastic, staying extended working hours for practice.

Production Deployment

We began the deployment on production on Saturday evening. This was after the OPD for the day got over. The deployment and setup finished by Sunday afternoon.

Rollout

Nikhil prepared a phase by phase rollout plan. In first phase they planned to use Diagnosis and Orders for a couple of months.

Senior Nurse using the system

They wanted a localised diagnosis list linked with ICD10 codes, such that doctors would enter in the words they are used to but they get reports with ICD10 codes. However they were finding difficult to prepare a list like that to upload to Bahmni. It was a lot of effort and required time from persons with clinical knowledge. One of the doctor mentioned that if a list like that is available it could be repurposed for many places in Indian subcontinent. I think we will put this on the roadmap for Bahmni. It will help in standardisation and faster rollouts.

Fixes — On-the-fly

When users started using the system, they started pointing out some issues. Some new requirements also came up. Amongst the ones that we could fix there, some required configuration work while some required minor tweaks to the product either in UI or in services like openerp-connect and openerp.

  • Discount order for doctors to communicate discounts to billing person (via auto-generated quotations)
  • A form and report based solution to manage surgery camps better.
  • Borders added to text boxes and buttons for better viewing on low-end screens, as can be seen in the above screenshot

We maintained a shared sheet of issues and requirements during the visit. The ones that we couldn’t fulfil there, we marked them as feature requirements.

After returning from our visit, we discussed the local fixes that we did onsite with the product team, who plan to merge these fixes into the product. We also shared the future requirements, which are being discussed and slotted in subsequent product releases.

Future Requirements

A few key requirements from SEARCH for the future are:

  • Offline Bahmni app to help them record data of a study they are doing in their community villages so that it helps physician in patient care on the field and allows program coordinator to monitor and course correct program with help of MIS reports. Also they would want data consumable by statistical software to derive insights from the study.
  • An App to help them in their community mobile clinics. This is not an urgent need though an area to be explored.
  • Ability to place mixed orders (Panels with lab tests across samples, Panels with lab investigations+radiology) to avoid clinical errors of missing out ordering some tests. Bahmni right now allows panels but only within samples and not across samples or types of orders.
  • Ability to uncheck tests from a selected lab panels while placing orders. Ordering by panel is convenient and reduces number of clicks but many time they would to skip one or two test from the panel. This would avoid incurring unnecessary time and money on when doctor start ordering by panels.
  • Ability to see cost of care to be able to easily decide on discounts to be given. Doctors many times decide the discount based on patients financial condition and total cost of care. They right now calculate themselves the cost of care in the current visit. Bahmni providing this would save them one mental cycle and probably utilise that in clinical care.
  • Alerts and reminders : One of the senior doctors mentioned that as such EMRs don’t necessarily provide immediate benefit to provider in patient care but alerts and reminders present potential of doing it. An example, he mentioned, is that many times providers forget to perform certain tests as decided by protocol or care plan. So if the provider is reminded of performing the test as per the original care plan, it could improve patient care.
  • Ability to place lab order for next visit. E.g. in Diabetes care, doctor orders in the current visit a fasting sugar test to be done when the patient comes for next visit. When the patient comes for the next visit, he has to directly go to the lab, get the tests done and then consult doctor. So doctors would like to place that order in the current visit itself.

Adoption

With doctors using the system, it means an important transition from operational efficiency to possibly improving direct patient care. However for a few, the question of effort of entering data vs. benefit still lingers.

Transition from paper and adoption of a new system by clinicians with the heavy patient load they operate in, is a key challenge at most implementations, which is true here as well.

For some the learning curve is high. One of the oldest doctor there needs assistance of a social worker for data entry.

The enthusiasm at the beginning was high and the phase by phase rollout would help in better adoption.

However when we checked the status of usage of the system after a couple of weeks, we learnt “When there is heavy patient load, they are finding it difficult to make data entry. So they have backlog”.


PACS / Digital X-ray integration

After multiple follow-ups, the digital X-ray engineers finally turned up on the last day of our visit. We needed them to enable the Network worklist feature. Since there was limited time left, we could only finish partial setup. I did the rest after coming back to Bangalore.


Towards the end of our visit there, rains had begun. Nikhil warned us to be careful to avoid stepping on a snake passing by.

Nikhil holding a non-poisonous snake

The schedule for us there turned out to be quite hectic, starting work morning 8.30am which continued till night 10pm daily with just breaks for lunch, dinner and tea. The hectic schedule continued even during our Hemalkasa visit. Read the next section to hear about that.


Lok Biradari Prakalp (LBP), Hemalkasa

A village health clinic inauguration at LBP

A brief about LBP

LBP was started as a social project by Baba Amte, a renowned social worker. The project has a hospital, a school and an animal orphanage.

A quote by Baba Amte that I really like :

I don’t want to be a leader. I want to be one who goes around with a little oil can and offer help when I see a breakdown.

The early years of the project entailed a massive struggle in the difficult conditions of a thick and remote forest. The centre, started in 1973, has recently developed into a full-fledged hospital of 40 beds and caters to over 45,000 patients annually, mainly tribals. For the interested reader, Pathways to Light is an inspiring autobiography by Dr. Prakash Amte, narrating stories about the LBP project.

Reason for Upgrade

LBP is a relatively light user of Bahmni. They use registration, lab and clinical modules. Data in clinical is entered by data entry operators, mainly to take out reports for funders and some studies.

Last deployment at LBP was done in June 2015. This upgrade was to provide some minor fixes, enhancements and integration with digital radiology and to ensure they are at the latest version of Bahmni so that it is easier to support.

Preparation & Planning

Preparation and planning for upgrade was similar to the one at SEARCH. The size of work, of course was much smaller. We planned for a week’s stay there.

On-the-site / In-the-field

We travelled from SEARCH to Hemalkasa by road, a journey of about 4 hours.

The distinctive fact about Hemalkasa is that during rainy season, it often gets completely cut-off from country land — broken road connectivity, no telephones and no internet. This could be a very challenging situation while upgrading, as Bahmni downloads installer from internet during installation.

To avoid getting into this situation, we had already prepared a cache of all the installers while in office. We used this cache to do the offline installation in training laptop and production.

Users and stakeholders were shown the new changes, feedback gathered, and changes made.

During the course of discussions and feedback solicitation, we also uncovered a few new requirements like the need to manage surgery camps (which we had already solved at SEARCH), follow up of patients, especially those with TB and the follow up for abnormal tests (similar to the ones I had seen while being at JSS). We setup forms and reports for it. They also wanted their sonography machine to be integrated such that they can fetch textual reports and store in EMR for easy viewing for subsequent visits of the patient.

I saw LBP’s new requests as a gradual move from solely operational efficiency to include direct patient care.


Amte’s Animal Ark

The wildlife orphanage and sanctuary is an interesting aspect of Lok Biradari Prakalp. I visited it multiple times during our one week stay over there. While the objective is to create awareness amongst tribals about these animals, I think it got created because of fearlessness and love of Dr. Prakash Amte for these animals.

Baba Amte, once talking about Prakash Amte and his love and fearlesness for animals, said

जिथे भिती असते तिथे प्रिती नसते

literally translating..

Where there is fear there is no love.
Me holding a 3 day old crocodile

By the time we returned heavy rainfalls had started in the region.


Things that worked well for us and we should continue doing

  1. Extensive analysis and multiple showcases before the visit.
  2. Timing of the visit based on weather.
  3. Having product code-base knowledge within the team that went for deployment.
  4. Having some buffer for unforeseen challenges.

When users start seeing a live system they always come up with more asks and suggestions. And since such places are remote without good internet connectivity, its always good to deliver as much as possible in a single visit. being prepared helps.


I like the experience of being part of executing such mini-projects end-to-end and the immediate gratification of doing something good that Bahmni implementations provide. They also provide opportunity to meet and learn from inspiring people and places. Thanks Bahmni!


Disclaimer: The views and opinions expressed by the author are personal to the author and do not necessarily reflect the positions of ThoughtWorks.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.