How Healthcare and Life Sciences Companies Fail: The Sequence a Board Can Interrupt

Corporate failure rarely begins with a cash shortfall. It begins in governance, moves through strategy and profit, and ends in liquidity. Artificial intelligence can shorten each stage, and the board has the most influence at the first.

In Brief

The Problem. Insolvency is the last stage of a sequence that starts years earlier: a crisis of governance and leadership, then of strategy, then of profits and cash flow, and finally of liquidity.

Why It Persists. The early signs are qualitative and easy to dismiss; the financial signs boards watch appear only after the options have narrowed. Artificial intelligence (AI) can shorten the time between them.

What Boards Should Do. Ask a defined question at each stage on a schedule, require independent validation of the claims the strategy rests on, and write the criteria for suspending a failing system into the original approval.

The Slow Evolution of Insolvency

Business insolvency can stop even the most promising enterprise, and healthcare, health technology, and life sciences companies have proved especially vulnerable. My argument is that insolvency is not an event but the final stage of a sequence, and that each stage is harder and costlier to reverse than the one before it. The board's influence is therefore greatest at the first stage, when the signs are least visible.

AI does not create this sequence, but it can compress it. A model can spread a flawed assumption across thousands of decisions before anyone reviews the results. A vendor commitment can lock in spending before anyone measures value. For each stage below, I describe a case, the practices that mitigate it, and the question a board should ask.

Stage One: The Crisis of Governance and Leadership

"The first responsibility of a leader is to define reality," Max De Pree wrote in Leadership Is an Art (Roberts, 2015). The first stage of insolvency begins when leadership stops defining reality accurately and the board does not notice. It shows up as a lack of strategic direction, poor oversight, and inadequate risk management. The board lacks independent information and the expertise to test management's claims, and it receives bad news slowly, if at all.

Theranos remains the clearest healthcare example. The company claimed a revolutionary blood-testing device. Its board included former Secretaries of State Henry Kissinger and George Shultz and former Senator Sam Nunn; critics pointed to its lack of medical expertise (Becker's Hospital Review, 2015). On January 3, 2022, a federal jury convicted founder Elizabeth Holmes of one count of conspiracy and three counts of wire fraud against investors. On July 7, 2022, a separate jury convicted former chief operating officer Ramesh Balwani on all 12 counts against him; the scheme targeted both investors and patients (U.S. Attorney's Office for the Northern District of California, 2022).

A board that cannot test management's claims about an AI system's accuracy stands where the Theranos board stood. In Marchand v. Barnhill, 212 A.3d 805 (Del. 2019), the Delaware Supreme Court held that directors must make a good-faith effort to put in place a board-level system for monitoring such risks (Duffy et al., 2019). That holding builds on In re Caremark International Inc. Derivative Litigation, 698 A.2d 959 (Del. Ch. 1996). AI used in clinical care, coverage decisions, or regulatory submissions is such a risk in most healthcare and life sciences organizations.

Practices that mitigate it. Adopt clear governance policies and define the roles of directors and executives. Build a culture of open communication, so that bad news reaches the board early. Invest in leadership development, including the board's own education.

The board question: who, independent of the team that built or bought the system, validated the performance claims we rely on?

Stage Two: The Crisis of Strategy

"The essence of strategy is choosing what not to do," Michael Porter wrote (Porter, 1996). A crisis of strategy follows when a company commits to a direction the evidence does not support or fails to adapt when markets change. In healthcare and life sciences, regulatory change and new technology often set it off.

IBM's Watson Health is the best-documented AI example. In July 2018, STAT reported that internal IBM documents showed Watson for Oncology had been trained on synthetic cases rather than real patient data and had recommended unsafe and incorrect cancer treatments (Ross & Swetlitz, 2018). In January 2022, IBM announced the sale of Watson Health's data and analytics assets to the private equity firm Francisco Partners (Ridley, 2022). The strategy had moved faster than the evidence for the capability at its center.

Practices that mitigate it. Review the strategic plan on a set schedule, not only when results disappoint. Test ideas before scaling them. Use data, including data on how AI performs in the organization's own setting, to inform strategic decisions. Bring in outside advisors when the strategy needs a view that management cannot supply.

The board question: which parts of our strategy rest on AI not yet proven in our own setting, and what evidence would change our course?

Stage Three: The Crisis of Profits and Cash Flow

A crisis of profits and cash flow emerges when expenses outpace revenue long enough to erode the margin of safety. Life sciences companies, with high research and development costs and long product cycles, are most exposed; health systems face the same pressure through thin operating margins.

Babylon Health shows how fast this stage can move, even with growing revenue. The digital health company went public in October 2021 through a USD 4.2 billion deal with a special purpose acquisition company (Landi, 2023b). Its revenue, driven by value-based care, rose from USD 323 million in 2021 to USD 1.1 billion in 2022, yet its loss widened from USD 83 million to USD 221 million. By March 31, 2023, it held USD 78 million in cash and was seeking interim funding while it planned to go private (Landi, 2023a). United Kingdom regulators and clinicians had questioned its corporate governance and patient safety as early as 2021 (Schwartz, 2023).

AI can speed this stage in two ways. Investments approved on projected savings may never be measured against those projections, and multi-year vendor commitments can fix costs while the benefits remain uncertain.

Practices that mitigate it. Maintain rigorous financial controls and reporting. Improve efficiency through process improvement and targeted cost reduction. Reassess the portfolio and divest what is not core. Seek turnaround advice early, while there is still time to use it.

The board question: for each AI investment above our approval threshold, what return did management project, and what has it measured since go-live?

Stage Four: The Crisis of Liquidity and Insolvency

The final stage arrives when a company cannot meet its short-term obligations. The choices narrow to restructuring, sale, or both, and control often passes to creditors.

Babylon reached that point within months. After a planned take-private deal collapsed, two of its U.S. subsidiaries filed for Chapter 7 liquidation in Delaware on August 9, 2023. On August 30, 2023, administrators were appointed for its United Kingdom holding company, and the United Kingdom business was sold to eMed Healthcare UK; shareholders received nothing (Landi, 2023b).

Steward Health Care shows the same stage in a hospital system. Steward, which operated 31 hospitals across eight states, filed for Chapter 11 in May 2024, the largest for-profit hospital bankruptcy in U.S. history, and several hospitals closed after receiving no qualified bids. Its former chief executive declined to testify before the Senate Committee on Health, Education, Labor, and Pensions, and the Senate adopted a criminal contempt referral; in August 2026, two senators asked the Department of Justice for an update (Mukherjee, 2026). The liquidity crisis drew public attention. The governance questions it raised began much earlier.

Practices that mitigate it. Prepare contingency plans for disruptions to cash flow. Communicate early with lenders, payers, and regulators, and negotiate before deadlines force terms. Engage restructuring advisors. Consider partnerships, mergers, or acquisitions that strengthen the balance sheet.

The board question: if our AI vendors or our own AI programs failed tomorrow, which obligations would we be unable to meet, and how quickly?

The Crisis Management Team

Successful mitigation depends on a formal crisis management team with a clear mandate. Its first task is to make sure the organization understands the crisis as it develops: to recognize the signs, state the problem plainly, and call attention to it. Its second is to secure the commitment of senior leaders, who must allocate resources and make difficult decisions. Its third is to bring individual initiatives under one plan, with objectives, timelines, and named owners. Its fourth is to protect the organization's people and culture through open communication and support for those affected.

For AI, the same team should own the decision to suspend or switch off a system. That decision is far easier to make when the criteria were written into the original approval.

The Strongest Objection: Companies Fail From Shocks

The strongest objection is that healthcare and life sciences companies fail mainly from external shocks: a reimbursement cut, a failed trial, a pandemic, a rise in interest rates. On this view, the sequence overstates what a board can control.

Shocks often set the timing of a failure. They do not set its outcome. Two organizations can face the same shock with very different options, depending on the information their boards received, the assumptions they tested, and the cash they preserved. Babylon's revenue was rising when its losses widened.

Where the Board Interrupts the Sequence

The board's influence declines at each stage, but it never disappears. At the governance stage, the board sets the information it receives. At the strategy stage, it tests assumptions. At the profit stage, it measures returns. At the liquidity stage, it chooses among narrowing options.

For the Board

• Does the board receive an inventory of AI systems in use, with a named accountable executive for each?
• Who independently validated the performance claims for our most consequential AI systems?
• Which strategic commitments depend on an AI capability not yet proven in our own setting?
• For each material AI investment, how do measured returns compare with the projections the board approved?
• What criteria would cause management to suspend an AI system or program, and where are those criteria recorded?
• Who would lead our crisis management team, and what would trigger its formation?

Receive new red flags and board questions each month in Oversight Rounds, from Avancer.

References

Becker's Hospital Review. (2015, October 29). Theranos trims board of directors from 12 to 5. https://www.beckershospitalreview.com/hospital-management-administration/theranos-trims-board-from-12-to-5-dismissing-henry-kissinger-and-others.html

Duffy, M. P., Ritts, G. J., & Heintz, A. F. (2019, August). Delaware Supreme Court reinforces directors' oversight obligations on mission-critical subjects. Jones Day. https://www.jonesday.com/en/insights/2019/08/delaware-court-reinforces-directors-oversight

Landi, H. (2023a, May 11). Digital health company Babylon Health plans to go private as losses grow. Fierce Healthcare. https://www.fiercehealthcare.com/digital-health/digital-health-company-babylon-health-plans-go-private-losses-grow

Landi, H. (2023b, September 5). Once a "unicorn," Babylon Health sells UK business to eMed via bankruptcy process. Fierce Healthcare. https://www.fiercehealthcare.com/digital-health/babylon-closes-us-business-lays-employees-after-mindmaze-take-private-deal-collapses

Mukherjee, S. (2026, August 11). Cassidy, Sanders press DOJ on former Steward CEO's contempt referral. Healthcare Dive. https://www.healthcaredive.com/news/cassidy-sanders-press-doj-former-steward-ceos-contempt-referral-ralph-de-la-torre/827566/

Porter, M. E. (1996). What is strategy? Harvard Business Review, 74(6). https://hbr.org/1996/11/what-is-strategy

Ridley, E. L. (2022, January 20). IBM to sell Watson Health assets to private equity firm. AuntMinnie. https://www.auntminnie.com/imaging-informatics/artificial-intelligence/article/15630230/ibm-to-sell-watson-health-assets-to-private-equity-firm

Roberts, M. D. (2015, April 18). God as the leader who defines reality. Max De Pree Center for Leadership. https://depree.org/life-for-leaders/god-as-the-leader-who-defines-reality/

Ross, C., & Swetlitz, I. (2018, July 25). IBM's Watson supercomputer recommended "unsafe and incorrect" cancer treatments, internal documents show. STAT. https://www.statnews.com/2018/07/25/ibm-watson-recommended-unsafe-incorrect-treatments/

Schwartz, N. (2023, August 31). How Babylon Health went from $2B to bankrupt. Becker's Hospital Review. https://beckershospitalreview.com/digital-health/how-babylon-health-went-from-2b-to-bankrupt.html

U.S. Attorney's Office for the Northern District of California. (2022, July 7). Theranos chief operating officer Ramesh "Sunny" Balwani found guilty of conspiracy, wire fraud [Press release]. U.S. Department of Justice. https://www.justice.gov/usao-ndca/pr/theranos-chief-operating-officer-ramesh-sunny-balwani-found-guilty-conspiracy-wire