Deaths in custody in prison facilities (2024–2026)

Introduction

This report constitutes an update and deepening of the analysis carried out by Cubalex in 2024 on deaths that occurred in the custody of the Cuban State.¹ That initial report aimed, from the outset, to analyze patterns, identify risk factors, and frame the events within the international standards applicable to deprivation of liberty, the use of force, and the state's obligation to prevent, investigate, and punish potentially unlawful deaths. This update builds upon those initial findings and projects them over time, with the objective of deepening the understanding of deaths occurring in state custody in Cuba between February 1, 2024, and June 30, 2026. In this first installment, we will focus our analysis on deaths linked to the prison context to identify aggravated manifestations of patterns that this report analyzes separately.

Regarding our conceptual framework, we rely on the definition of potentially unlawful deaths as a broad legal category developed in international human rights law to identify deaths in which there is reasonable suspicion of a violation of the right to life. According to the Minnesota Protocol on the Investigation of Potentially Unlawful Deaths (2016), this category encompasses not only deaths caused directly by state agents, but also those in which the state may have contributed through action, acquiescence, or omission, including situations in which it failed to fulfill its duty to prevent foreseeable risks to life. The Protocol establishes that the duty to investigate is triggered whenever the state has knowledge, or should have had knowledge, of a potentially unlawful death, even in the absence of a formal complaint, thus reinforcing the preventive and protective nature of this international obligation.²

Within this concept, deaths in custody are classified as a specific subcategory characterized by the direct and effective control of the State over the deceased. The Minnesota Protocol recognizes that when a person dies while deprived of their liberty or in the custody of state authorities, a general presumption of state responsibility is triggered. This presumption is based on the enhanced duty of care incumbent upon the State in contexts of deprivation of liberty, where the person is entirely dependent on the authorities for their safety, integrity, and access to essential services such as medical care.³

For the purposes of this report, custody is defined as the exercise of effective state control over a person's life, integrity, and liberty. This material and functional concept is supported by General Comment No. 35 of the Human Rights Committee, which expressly includes involuntary hospitalization and transfer against one's will; by the jurisprudence of the Inter-American Court of Human Rights on the special relationship of subjection and the State's position as guarantor with respect to any person under its control, authority, or custody (Ximenes Lopes v. Brazil, 2006; Vélez Loor v. Panama, 2010; Chinchilla Sandoval et al. v. Guatemala, 2016); and by the United Nations Standard Minimum Rules for the Treatment of Prisoners (Mandela Rules), Rule 27 of which establishes that transfer to an external health facility does not interrupt the status of custody or the responsibility of the authorities. In practice, the continued existence of custody is verified through three operational indicators: continuous or delegated state surveillance; objective restriction of movement, including the inability of the person or their family to decide on discharge or transfer; and the subordination of discharge or destination to the authorization of a police, judicial, or administrative authority. The presence of any of these elements constitutes material custody, regardless of the legal classification that the State itself assigns to the situation.⁴

In this sense, it is legally consistent to assert that every death in custody should be considered, at first glance, as a potentially unlawful death. This does not imply that all deaths in custody are automatically unlawful, but rather that, given the conditions of absolute state control, there is an initial suspicion that compels a rigorous, independent, and effective investigation. The burden of proof, in these cases, shifts to the state, which must demonstrate that it fulfilled its obligations of protection and that the death was not the result of an action or omission attributable to its agents.

Regarding the relationship with extrajudicial executions, these should be understood as an aggravated form of potentially unlawful killing, characterized by the deprivation of life as a result of the direct action of state agents—or private individuals acting with their authorization, support, or acquiescence—through the use of lethal or potentially lethal force in violation of legal and human rights standards. Unlike other potentially unlawful killings, which may result from state omissions, negligence, or lack of protection, extrajudicial executions are primarily linked to active conduct that generates a foreseeable risk of death, such as shootings, beatings, asphyxiation, ramming, or any form of violence incompatible with the legitimate use of force.

While some extrajudicial killings may occur in custodial settings, many take place outside of them, for example, during police operations, criminal prosecutions, or immigration enforcement. However, in contexts of deprivation of liberty, the distinction between death in custody and extrajudicial killing is not always mutually exclusive, since a single death can meet both criteria when there is evidence that state agents directly caused the death through the use of force. In these cases, the classification does not depend on proof of a specific intent to kill—which is often inaccessible in contexts of institutional opacity—but rather on the existence of objective elements that allow one to infer that the death was the consequence of an arbitrary, abusive, or incompatible state action that violates the duty to guarantee life.

In this regard, according to the Minnesota Protocol, the analysis should focus on whether the death could have been caused by acts or omissions of the State or in violation of its obligation to protect life, and not solely on the subjective intent of the perpetrator. This approach allows for a more appropriate classification in contexts where access to information is limited, investigations lack independence, and patterns of cover-up exist. Thus, the identification of a possible extrajudicial execution can be based on indicators such as the presence of injuries inconsistent with the official version, the lack of documentation, the victim's prior incommunicado status, or contradictions in the State's account, without it being necessary to fully establish the perpetrator's intent at this stage.

Consequently, the relationship between these categories can be understood as follows: potentially unlawful death functions as an umbrella category encompassing all deaths for which there are indications of state responsibility, whether by action, omission, or failure to fulfill the duty of protection; death in custody is defined by the context of state control over the person and activates a reinforced presumption of responsibility; and extrajudicial execution is configured by the type of conduct attributable to state agents, particularly when death results from the use of lethal or potentially lethal force, or arbitrary actions incompatible with the standards governing the use of force.

In this context, the classification does not depend solely on proof of a specific intent to kill, but rather on the existence of objective elements that allow one to infer that the deprivation of life is a consequence of abusive state action or action contrary to its international obligations. This distinction allows for a coherent analysis of cases in contexts of high opacity, without losing sight of the fact that all of them share a common obligation of the State: to respect, protect, and guarantee the right to life in accordance with international standards.

Main mechanisms of death analyzed in the prison context

For the purposes of this report, the mechanism of death is defined as the process by which certain actions, omissions, or conditions of confinement lead to or contribute to a person's death. The term is used analytically to examine how the outcome occurred and its relationship to the actions of the authorities, the conditions of custody, and the fulfillment of the state's duty of protection. The classification is based on the information available for each case and allows for the identification of recurring patterns, without replacing the investigation necessary to clarify the circumstances and determine the corresponding responsibilities.

The first mechanism analyzed is structural omission, which includes deaths linked to negligence, denial, or delay of medical care, as well as state failures in providing basic conditions of confinement, especially food, water, and sanitation. Within this mechanism, four sub-patterns are distinguished: deaths from malnutrition or starvation; those associated with communicable diseases, particularly tuberculosis; those occurring during hunger strikes in contexts of neglect of risk; and those linked to dehydration. The common element is the lack of an adequate institutional response to essential needs or life-threatening situations.

The second mechanism is lethal direct action, which encompasses deaths attributed to state violence while in custody and those where there is evidence that requires further investigation into possible state intervention. Within this category are identified cases classified as extrajudicial executions, which include beatings resulting in death and the use of lethal force by state agents. According to the classification criteria used in the report, cases are also included where the information gathered expressly indicates physical violence as the cause of death, even if the perpetrator's identity or status is not fully confirmed.

Within the category of lethal direct action are also deaths of particular suspicion, in which an official version of suicide coexists with evidence of violence by state agents, testimonies that challenge the official suicide version, or indications of possible cover-ups by the authorities, without it being possible to definitively identify the perpetrator. In these cases, their inclusion stems from the need to examine the possible involvement of state violence under the enhanced standard of investigation and the presumption of state responsibility applicable to violent deaths in custody, in accordance with the Minnesota Protocol, without presenting as clear the authorship that remains to be determined.

The third mechanism corresponds to the allegation of suicide in custody. This includes deaths attributed to self-harm where no consistent evidence of direct state violence has been identified as the immediate cause of death. This classification does not exclude the existence of a history of abuse, neglect of mental health, or failure to provide protection that could have increased the risk of suicide.

The fourth mechanism is violence among inmates, which includes deaths caused by other prisoners. Its analysis is not limited to identifying the immediate perpetrator of the aggression, but also examines the state's fulfillment of its duty to protect third parties, particularly the measures of prevention, control, and supervision within prison facilities. The involvement of another inmate does not, therefore, preclude the state's potential responsibility for actions or omissions that may have contributed to the outcome.

The report examines the release from prison in a differentiated way. in extremis This category is considered cross-cutting, not constituting an independent mechanism of death. It refers to releases or extra-penal licenses granted when the process leading to death, originating during incarceration, was already irreversible and the death occurred hours or days later. The seven cases identified within the universe of 115 deaths are counted according to the corresponding mechanism—mostly structural omission—without being added back to the total. Their analysis as a pattern of externalizing death considers the formal release from custody in the terminal phase and the risk that the death will be excluded from records limited to deaths formally occurring in custody, without this release interrupting the causal link with the conditions of incarceration.

Finally, the "mechanism not determined" category includes cases where the available information is insufficient to establish how the death occurred. This includes records that do not specify the cause, as well as those that only mention a heart attack or cardiac arrest, without providing any evidence of negligence, violence, or cover-up. This category reflects a limitation in the information gathered, not a conclusion about the absence of state responsibility, and does not relieve the state of its obligation to investigate and clarify the circumstances of the death.

1. Delimitation of the universe

Between February 1, 2024 and June 30, 2026, Cubalex recorded 115 deaths in state custody in penitentiary establishments, corresponding to people deprived of liberty by judicial resolution.⁵ This universe includes seven cases of people who were granted release from prison or extra-penal license in extremisThese are situations in which release occurred when the process leading to death—which originated decisively during incarceration—was already irreversible, and death occurred within hours or days of formal release from prison. Their inclusion stems from the fact that, in these cases, release does not break the causal link between the conditions of custody and the fatal outcome, nor does it extinguish the State's responsibility as guarantor. This subgroup is analyzed as a specific pattern in section 2.5. Because it is a cross-cutting circumstance and not a mechanism of death, these seven cases are counted according to their actual mechanism—mostly structural omission—and not as a separate category added to the total.

2. Main mechanisms of death in custody

Analysis of incidents linked to death reveals that lethality in prison custody does not follow a homogeneous pattern, but rather is organized around four structural mechanisms: structural omission; direct lethal action—which includes extrajudicial executions and deaths under special suspicion; the claim of suicide; and violence among inmates with a breach of the duty of protection. Release in extremis is addressed separately, as a cross-cutting pattern of externalizing death (section 2.5): it does not constitute an autonomous mechanism, since these deaths are accounted for according to their actual mechanism, mostly structural omission. In eight cases, the available information did not allow for a precise classification of the mechanism, which does not diminish their relevance within the overall documented pattern nor relieve the State of its obligation to investigate them.

2.1 Structural omission (negligence or denial of medical care)

Denial of medical care or negligence is the dominant mechanism, identified in at least 80 of the 115 recorded deaths (≈70%). This figure includes the seven cases of last-minute release (section 2.5), where the mechanism of death was omission, and therefore these are counted here and not as a separate category. This pattern manifests itself systematically and particularly affects people in situations of greater vulnerability. It is a mechanism of death linked to failures in the provision of medical care, delays in assistance, and omissions in contexts of known risk. Its consistency, territorial distribution, and association with vulnerability variables allow it to be interpreted as a form of institutional violence by omission, embedded in the regular functioning of the prison system.

The documented cases illustrate the extreme seriousness of these omissions: a 59-year-old inmate waited more than 12 hours at the medical post of the El Típico Nuevo prison (Las Tunas) to be transferred to a hospital after suffering a stroke; an elderly cancer patient remained for about a month with a catheter without a collection bag in the Típico Viejo prison until he died;⁶ An inmate diagnosed with epilepsy in Boniato was left lying on the floor, without assistance or supervision, after suffering another seizure;⁷ And a 69-year-old inmate in Granma, beaten a month earlier with a board studded with nails, died after his injuries and the resulting fracture worsened without medical attention. Within this context, the volume and recurrence of cases allow us to identify two sub-patterns that require differentiated treatment.

2.1.1 Deaths due to malnutrition and starvation.

At least 21 deaths were documented that were directly attributed to severe malnutrition or starvation, including cases of metabolic coma due to starvation and grade 3 malnutrition without medical attention or dietary modification.

The very existence, in Boniato prison, of wards designated for “malnourished” or “underweight” individuals—where several of these deaths were recorded—constitutes an institutional acknowledgment of the magnitude of the problem and its foreseeable nature. This pattern transcends the category of individual medical negligence, as it is not merely a matter of omitting assistance in the face of illness, but rather a failure of the State to provide basic nutrition, a minimum and non-delegable condition for any deprivation of liberty according to the United Nations Standard Minimum Rules for the Treatment of Prisoners (Nelson Mandela Rules).⁸

Among the victims is political prisoner Ernesto Brieva Sempé, who died in the Combinado del Este prison in a context of severe malnutrition and kidney complications that, according to his family, were not adequately treated.⁹ The concentration of these cases in certain facilities—particularly the Guantánamo Combined Camp, where at least five deaths associated with malnutrition or starvation were documented, mostly during 2026, and the Boniato prison—indicates acute food crises in specific centers, with foreseeable and preventable lethal outcomes. This sub-pattern is also associated with three deaths that occurred during hunger strikes, including that of political prisoner Yan Carlos González González, who had been on hunger strike for months in protest against his trial and sentence.,¹⁰ without the authorities taking effective measures to protect their lives.¹¹

2.1.2 Deaths from communicable diseases and collective health conditions.

Twelve deaths linked to tuberculosis were documented, with concentrations that reveal the collective nature of the risk: three deaths at the Agüica prison (Matanzas) during the same period, one at the La Condesa prison (Mayabeque), and several cases associated with the Boniato prison, in addition to an inmate who reportedly contracted the disease within the prison in Santiago de Cuba and was transferred to the hospital in a terminal state, with critically low hemoglobin levels. The recurrence of tuberculosis-related deaths in the same facilities raises the need to clarify the conditions of detection, treatment, sanitary isolation, and prevention of contagion, as well as the possible impact of overcrowding and other material deficiencies.¹²

2.2 Lethal direct action

Nine cases of extrajudicial killings (≈8%) were identified, characterized by the direct use of force by state agents or by evidence of physical violence cited as the cause of death, even when the perpetrator is not fully confirmed. These include documented beatings resulting in death inflicted by prison guards—such as those that cost the life of an inmate at Melena 2 prison and a 73-year-old inmate at Kilo 8 prison in Camagüey.¹³ already a common prisoner in the Combinado del Este prison—, the case of political prisoner Manuel de Jesús Guillén Esplugas, beaten by guards at the Combinado del Este prison after being recaptured following an escape and whose death was officially presented as a suicide,¹⁴ and the death of an inmate from detachment 18 of the Boniato prison at the hands of the guards during a nighttime escape attempt.

The use of lethal force against an escape that did not represent an imminent threat to the life of third parties is incompatible with the Basic Principles on the Use of Force and Firearms by Law Enforcement Officials, which only permit the intentional use of lethal weapons when strictly unavoidable to protect a life.¹⁵ Although numerically less frequent, these deaths represent the most serious level of violation of the right to life within the prison system, and the coexistence, in several cases, of official narratives of suicide or hanging incompatible with empirical evidence demonstrates the operation of institutional cover-up mechanisms.

Within the category of lethal direct action, a distinct subset includes three particularly suspicious deaths in which the official suicide narrative coexists with significant evidence of prior or concurrent state violence. These deaths are counted separately from executions—bringing the total to 12 cases of lethal direct action—but are grouped under this category because they trigger the presumption of state responsibility for all violent deaths in custody and the enhanced duty to investigate under the Minnesota Protocol. At the Bungo 8 prison camp in Santiago de Cuba, authorities claimed that an inmate had committed suicide, but the body showed multiple bruises, and there was no security presence or explanation at the funeral home, fueling the family's accusation that the death, following a beating, was covered up as a suicide.¹⁶ In the Guatao women's prison (Havana), a 33-year-old inmate was found dead in a cell hours after an altercation with staff at the center; testimonies indicate that the body had marks on the neck and that it was transferred to forensic medicine without the presence of family members, who were initially not allowed to see it.¹⁷ And in the Canaleta prison (Ciego de Ávila), a 20-year-old inmate who had staged a hunger protest was beaten and taken to a punishment cell, after which the authorities reported that he had hanged himself; inmates who documented the moment maintain that he may have been strangled by the authorities themselves.¹⁸

This last case occurred in the context of a protest within the prison, introducing a specific risk scenario: the repressive response to acts of protest by the prison population. According to the methodology of this report, these cases are excluded from the suicide count and must be treated under the enhanced investigative standard applicable to possible extrajudicial executions with simulation. Along the same lines of institutional violence, the death of an inmate in the Guantánamo provincial prison, who was kept restrained for three days in “mechanical restraint” until he died, should be mentioned.¹⁹The prolonged use of coercive measures against a person constitutes a form of cruel treatment inconsistent with the Nelson Mandela Rules.²⁰

2.3 Allegation of suicide in custody

Thirteen deaths were recorded and attributed to suicide in the collected information. Cases where available evidence cast doubt on whether the death was self-inflicted were excluded from this subset, as were those with documented histories of violence, degrading treatment, or neglect that could have increased the risk of suicide. This subset raises serious questions about detention conditions, institutional supervision, access to mental health care, and compliance with the duty of prevention, since the records themselves document known risk factors that were ignored by the authorities: one inmate at Boniato prison committed suicide in a context of extreme malnutrition and daily fainting spells, having gone approximately a month without receiving medical attention; another suffered from depression that prison authorities failed to address, also refusing to transfer him to a different unit when he requested it;²¹ A third inmate took his own life at Las Mangas prison despite a previous suicide attempt for which he received no psychological support; a 60-year-old woman died by hanging at the Camagüey women's prison, where Cubalex received information that she was publicly and constantly humiliated by the unit's head officer; and an inmate at the Guantánamo Combined Prison hanged himself after being beaten by prisoners working for the garrison. In contexts of deprivation of liberty, the State must not only refrain from causing death, but also actively prevent situations of self-harm, including those generated by degrading institutional treatment itself.²² These cases should therefore be interpreted as indicators of structural failures in the systems of custody, supervision and protection of mental health.

2.4 Violence among inmates and failure to fulfill the duty of protection

Two deaths caused by other inmates were documented in contexts that demonstrate the State's failure to fulfill its duty to protect the lives of persons in custody from third parties: a 56-year-old inmate killed by another prisoner at Canaletas Prison (Perico, Matanzas), and an elderly inmate stabbed to death by two other prisoners in the infirmary of Ariza Prison (Cienfuegos), allegedly over a piece of bread. According to the Minnesota Protocol and Inter-American jurisprudence, the State, in its position as guarantor, is also responsible for deaths caused by individuals under its custody when it has failed to fulfill its duty of prevention, control, and supervision.²³ The second case is particularly revealing of the interaction between this axis and the structural failures described: a lethal dispute over food, inside the infirmary of a prison, condenses in a single event the food shortage, the lack of protection of the most vulnerable people and the absence of effective supervision even in health care spaces.

2.5 Release in extremis and externalization of death

The monitoring identified a specific pattern that warrants differentiated treatment: the granting of early release or extra-penal leave when the death of the inmate is imminent and irreversible. Seven cases were documented. Early release is not a mechanism of death, but rather a cross-cutting circumstance: these seven cases are not counted as a separate category, but are tallied under their actual mechanism—mostly structural omission (section 2.1)—and are analyzed here solely as a pattern of externalizing death. Among them: an inmate at Boniato prison who was granted extra-penal leave due to his poor health and died at home the following day, after having spent 28 days without receiving medical attention in prison;²⁴ An inmate transferred from Mar Verde and Boniato to the hospital with signs of heart, kidney and tuberculosis conditions, to whom extra-penal release was applied two days before his death in the prison ward; an inmate from the Guamajal prison whose extra-penal release was approved when he was already in the intensive care unit, dying days later with diabetes, hypertension and malnutrition; a man with paranoid schizophrenia and multiple chronic conditions “released” at 8:00 a.m. and died at 2:00 p.m. that same day in a Havana hospital, whose body showed blows to the chest as confirmed by his family;²⁵ an inmate from Boniato released when his death was imminent; a prisoner transferred from Quivicán to the hospital “malnourished, covered in blows and bruises, with an extra-penal license for one month”, according to his family;²⁶ and the political prisoner of July 11, Luis Miguel Oña Jiménez, an HIV patient who had suffered an ischemia in the Cuba-Panama prison, who was granted extra-penal license and sent home when his health condition was irreversible, dying days later.²⁷

This practice, in effect, functions as a mechanism for externalizing death: the person does not formally die “in prison,” which artificially reduces prison mortality statistics, obscures the traceability of state responsibility, and shifts the burden of care—which the state failed to provide throughout the incarceration—to families in the terminal phase. The pattern is exacerbated by its opposite: the denial or delay of early release when it could have saved a life, as in the case of a 52-year-old inmate at Veguitas 1 prison who had been waiting for ten months for early release while the authorities failed to guarantee him the medication he needed, and who died before the measure was implemented.

The proximity between release and death necessitates examining the appropriateness of the measure and the conditions of care during incarceration. These situations can hinder the traceability of the death and facilitate its exclusion from records limited to deaths formally occurring in custody. Release from the facility does not eliminate the need to investigate prior acts or omissions that may have contributed to the outcome. Based on the Minnesota Protocol standard, deaths occurring after these releases retain their potentially unlawful nature, as the lethal process originated and was consolidated while in state custody.

3. Territorial distribution and differentiated configurations

The territorial analysis reveals a marked geographic concentration of prison lethality, with three provinces accounting for more than half of the cases: Santiago de Cuba (29 deaths, 25.2%), Havana (21 deaths, 18.3%), and Mayabeque (16 deaths, 13.9%). Guantánamo follows, consolidating its position as the fourth hotspot with 11 deaths (9.6%), most of them concentrated in 2026 and associated with malnutrition, starvation, and tuberculosis, and Granma with 7 deaths (6.1%). This pattern indicates that deaths in custody are clustered in certain jurisdictions, suggesting the existence of differentiated institutional dynamics at the territorial level, with direct implications for the attribution of state responsibility and the design of prevention measures.

At the municipal level, this concentration intensifies even more markedly. The municipality of Santiago de Cuba alone accounts for 28 deaths (24.31% of the total), followed by Guantánamo with 11 (9.61%), Guanabacoa (Havana) with 11 (9.61%), and Quivicán (Mayabeque) with 7 (6.11%), plus one additional case linked to the Quivicán prison registered in a neighboring municipality. Based on the documented information, it is possible to identify the prisons with the highest number of these incidents: Boniato Prison (Santiago de Cuba), Combinado del Este (Havana), Combinado de Guantánamo, Quivicán Prison (Mayabeque), and Las Mangas Prison (Granma), along with secondary outbreaks in La Condesa and Melena prisons (Mayabeque), Ariza (Cienfuegos), and Canaleta (Ciego de Ávila). The Agüica prison (Matanzas) deserves special mention for its concentration of tuberculosis deaths, even though its total number is lower. This distribution reveals highly localized hotspots of mortality, where the risk of death in custody is substantially higher than in other areas of the country.

The combined examination of territorial distribution and death mechanisms reveals distinct patterns. Guantánamo and Matanzas are characterized by acute structural neglect, with a predominance of deaths due to malnutrition and communicable diseases, indicative of food and health crises in specific facilities. Santiago de Cuba combines a high volume of deaths with the coexistence of negligence, malnutrition, allegations of suicide, and direct lethal action, indicating the concurrence of multiple risk factors in the same territory, with Boniato as its epicenter. Havana—particularly Guanabacoa, home to the Combinado del Este prison—presents a mixed pattern in which medical negligence and forms of direct lethal violence coexist, including the death from malnutrition of a political prisoner and extrajudicial executions presented as suicides. Mayabeque exhibits a clear predominance of structural neglect, with a high proportion of deaths associated with medical negligence and tuberculosis in the Quivicán, La Condesa, and Melena prisons. Ciego de Ávila, for its part, emerges as a scene of institutional violence associated with protests within the Canaleta prison.

4. Intersectional dimension: vulnerability and differential risk

The analysis incorporates an intersectional dimension that allows for the identification of groups particularly vulnerable to certain mechanisms of death in custody. This approach is key to understanding that prison lethality does not affect the incarcerated population uniformly, but rather is distributed unevenly according to specific conditions of vulnerability, which interact with the structural shortcomings of the system.

4.1 Links to independent civil society and detention for political reasons

The records identify nine victims as members of independent civil society organizations and political prisoners, alluding to the political motivation behind their detention, with several cases linked to the July 11 protests. Among them are Manuel de Jesús Guillén Esplugas—associated with UNPACU and Cuba Decide, beaten by guards at Combinado del Este prison and whose death was presented as a suicide—, and Lázaro García Ríos, returned from the hospital to Combinado del Este without having recovered and who died in the prison;²⁸ Yan Carlos González González, who died during a hunger strike; political prisoner Yoleisy Oviedo Rodríguez, who died in the camp near El Guatao, amid allegations of medical neglect, without the causes of her death being clarified; her family was initially unable to access the body;²⁹ Ernesto Brieva Sempé, who died from severe malnutrition and kidney complications at the Combinado del Este;³⁰ and Luis Miguel Oña Jiménez, an HIV patient who died days after receiving an extra-penal license in extremis. The fact that the deaths of people imprisoned for political reasons are repeatedly associated with the neglect of basic needs such as food and healthcare, direct violence, and cover-ups reinforces concerns about the unequal treatment they receive within the prison system. This information should be interpreted with caution, as it only reflects data collected in the database and does not rule out underreporting.

4.2 Chronic illness, age and mental health

Twelve individuals with chronic illnesses—including cancer patients, those with HIV, and those with epilepsy—were identified, all of whom died from negligence, demonstrating a complete correlation between this condition and the mechanism of medical omission. This pattern is replicated among older adults: at least seven deaths of individuals aged 60 and over were documented in contexts of neglect, malnutrition, or violence with lack of protection, including a 73-year-old inmate who died after being beaten by state agents in Camagüey and an elderly person murdered in the infirmary of the Ariza prison. Furthermore, at least three victims were identified as psychiatric patients, and in several cases of alleged suicide, records document untreated depression or previous attempts without psychological follow-up. The convergence of these factors allows for the identification of a clearly differentiated group: individuals with greater healthcare and mental health needs who, instead of receiving the enhanced protection their condition requires, are disproportionately exposed to institutional neglect.³¹

4.3 Racial bias

People of African descent represent 27.1% of the analyzed prison population (31 out of 115 cases), and a clear majority of them—21 cases, equivalent to 67.7%—died in contexts of medical negligence, malnutrition, or communicable diseases, compared to a significantly lower proportion in the rest of the population. This difference suggests a racialized sub-pattern in the way deaths in custody occur, particularly in relation to access to food and healthcare. The relative overexposure of people of African descent to mechanisms of omission is incompatible with the claim of equal treatment in access to basic services within the prison system and constitutes an additional factor of vulnerability within an already restrictive environment. It is also worth noting that the three women who died in the analyzed population were of African descent.

4.4 Sociodemographic and gender dimension

The analyzed population comprises 112 men and 3 women, confirming that incarceration lethality occurs within a heavily male-dominated context, consistent with the composition of the prison system. Although women represent a small proportion of the documented deaths, the three cases present serious circumstances requiring specific attention: allegations of neglect of healthcare, evidence of violence, a history of degrading treatment, and restrictions on information or access for families. Their analysis demands a gender perspective, and the number of cases does not allow for establishing a greater relative severity compared to the overall number of victims.

Political prisoner Yoleisy Oviedo Rodríguez died amid allegations of medical neglect despite pre-existing health conditions, and her body was initially withheld from her family. A 33-year-old inmate was found dead in a cell at the Guatao women's prison, with marks on her neck, hours after an altercation with prison staff. A 60-year-old woman died by hanging at the Camagüey women's prison after being subjected to repeated public humiliation by an officer. Two of these three deaths are linked to the Guatao prison complex. The convergence, within such a small subset, of allegations of suicide, a history of violence or degrading institutional treatment, and a lack of transparency toward families suggests that the low statistical representation of women may be obscuring specific dynamics of institutional violence in women's prisons, which require differentiated monitoring.³²

Regarding age, information was available for 43 victims. Based on this data, the overall average age is 41.9 years, the median is 42, and the range is between 20 and 73 years. This profile indicates that in-custody lethality is not concentrated exclusively at the extremes of age, but significantly affects working-age adults, while also registering both very young victims—between 20 and 23 years old, associated with contexts of violence and political imprisonment—and older individuals with aggravated vulnerability. Consequently, the explanatory factors of this phenomenon should be located primarily in the structural conditions of detention and the institutional functioning of the prison system, rather than in individual characteristics associated with age.

5. Interpretive synthesis

The analysis of documented deaths reveals recurring patterns of neglect, violence, and lack of protection that compromise the right to life in prisons. Structural omission is the predominant factor: negligence or denial of medical care was identified in at least 80 of the 115 cases, approximately 70% of the analyzed population. Its recurrence demonstrates that the problem extends beyond individual care failures and demands an examination of the organization and operation of prison services. Deaths associated with malnutrition and starvation at the Guantánamo and Boniato prisons, along with the concentration of tuberculosis-related deaths at Agüica and Boniato, reinforce concerns about serious deficiencies in nutrition, healthcare, and the prevention of collective risks.

Direct lethal violence constitutes another central dimension of the phenomenon. The cases described include beatings attributed to prison guards and the use of lethal force during an escape attempt. These incidents are compounded by the prolonged use of coercive measures and suspicious deaths in which official suicide narratives coexist with evidence of violence, even in the context of prison protests. Within the framework of direct lethal action, suspicious deaths—official suicide narratives contradicted by evidence of violence—require the application of a heightened standard of investigation and the presumption of state responsibility, without automatically accepting the official explanation. Cases whose mechanism remains undetermined, on the other hand, are deaths for which the available information does not allow for a definitive cause to be established; this does not exempt the State from investigating them until the circumstances are clarified.

Deaths attributed to suicide and those caused by other inmates highlight the importance of prevention and protection duties. Histories of untreated depression, previous suicide attempts without follow-up, and degrading treatment preclude analyzing the former cases in isolation from custody conditions and the institutional response to known risks. Similarly, violence among inmates necessitates examining supervision and protection measures against third parties, as well as the potential impact of shortages and inadequate protection. Neither attributing a death to suicide nor the involvement of another inmate precludes examining state actions or omissions that may have contributed to the outcome.

The releases in extremis They add a specific dimension to the analysis of the institutional response. The granting of non-penal licenses in the period immediately following death, along with cases of denial or delay during a decline in health, raises doubts about the appropriateness and protective effectiveness of this mechanism. Formal release from prison does not eliminate the need to investigate the conditions of confinement or potential liability for prior acts or omissions. It can also hinder the traceability of the death and facilitate its exclusion from records limited to deaths formally occurring in custody, although the report does not allow us to conclude that this exclusion has actually taken place in official statistics.

The territorial distribution also reveals the repetition of certain mechanisms in specific facilities and the concurrence of different forms of vulnerability in others. These configurations guide the examination of institutional practices and the identification of prevention priorities. The intersectional dimension allows us to examine how custodial conditions respond—or fail to respond—to specific needs related to illness, age, and mental health. Cases involving women, people of African descent, and people imprisoned for political reasons also warrant differentiated monitoring, attentive to potential forms of discrimination and violence.

Taken together, the findings point to structural patterns and causes in the documented deaths: the recurrence of medical omissions, material deprivation, violence, and failures to provide protection reveal problems that extend beyond the individual actions of certain officials. Clarifying each death must be accompanied by an examination of the institutional practices and conditions that contribute to this lack of protection. The limitations of the registry do not diminish the gravity of the events or the need to investigate, determine responsibility, and adopt measures to guarantee the lives of people deprived of their liberty.

6. Recommendations

Based on the documented patterns, Cubalex makes the following recommendations to the Cuban State, aimed at preventing further deaths, clarifying the recorded deaths, and correcting the conditions that compromise the life and integrity of people deprived of their liberty.

I. Guarantee timely healthcare and continuity of treatment

Ensure effective access to healthcare personnel, medications, diagnostic tests, and specialized care, with decisions based on clinical criteria and free from disciplinary or security interference that unjustifiably delays care. Establish verifiable procedures for identifying emergencies, immediate transfer to hospitals, and follow-up on medical orders, preventing return to the facility when it cannot guarantee the necessary care.

Strengthen the detection and treatment of communicable diseases, especially tuberculosis, through timely assessments, continuity of care, and infection prevention measures. Sanitary isolation should be used exclusively for clinical needs and not as a form of punishment. Furthermore, guarantee independent medical care for individuals on hunger strike, respecting their autonomy and informed consent, and ensuring that their protest does not justify denying them assistance.

II. Prevent malnutrition and correct the material conditions that aggravate health risks

Ensure sufficient, high-quality food that meets health needs, as well as permanent access to safe drinking water and appropriate hygiene and ventilation. Meeting these needs should not depend on the resources or capacity of families to provide assistance.

Take urgent measures in facilities where deaths from malnutrition or starvation have been documented, including nutritional assessments, individualized care, and health monitoring of food quantity and quality. Identifying underweight individuals should trigger an effective protection and treatment response, not simply classify or group them within the prison. Also examine the potential impact of overcrowding and other material deficiencies on disease transmission and health deterioration.

III. Preventing violence by prison officers and protecting against attacks by other inmates

Eradicate beatings, corporal punishment, and any intervention intended to intimidate, humiliate, or punish through violence. Ensure the use of force adheres to criteria of legality, strict necessity, and proportionality, and prevent an escape or protest from constituting, in and of itself, a justification for the use of lethal force.

Prohibit the use of coercive measures as punishment and limit any permitted use to the strictly necessary time, with recording, supervision, and evaluation of their effects on health. Strengthen the prevention of aggression among inmates by identifying threats, protecting those at risk, and effectively supervising living spaces and healthcare facilities, without delegating control or punishment functions to other inmates.

IV. Establish an effective response to the risk of suicide and self-harm

Implement procedures for detecting and monitoring suicide risk from admission and throughout incarceration, paying particular attention to a history of self-harm, depression, physical deterioration, threats, and degrading treatment. Every previous attempt should result in a clinical evaluation and a care and protection plan, not just surveillance measures.

Ensure timely access to psychological and psychiatric care and review living conditions when they may exacerbate the risk. Prevention should not rely on punitive isolation or humiliating practices. Also investigate the possible contribution of institutional treatment and failure to provide assistance to deaths attributed to suicide.

V. Ensure the timely evaluation of extra-penal licenses and other measures of release from prison for health reasons

Review the procedures for assessing and granting these measures so they can serve a humanitarian purpose before the deterioration becomes irreversible. Establish response timelines, reasoned decisions, access to independent medical assessments, and effective mechanisms for reviewing denials or delays. The assessment should consider both the individual's health condition and the facility's actual capacity to provide the necessary treatment and care.

Maintain traceability of subsequent deaths when there are indications of a link to the conditions of confinement and investigate acts or omissions prior to release from prison.

VI. Independently investigate each death and determine the corresponding responsibilities

Ensure prompt, thorough, impartial, and transparent investigations into documented deaths, including those attributed to suicide, those caused by other inmates, and those occurring after releases from prison. in extremis, In accordance with the standards of the Minnesota Protocol, preserve evidence, ensure the independence of medico-legal examinations, and facilitate access to clinical records, custody records, and relevant testimonies.

When official accounts are inconsistent with injuries, testimonies, or other evidence, investigate both the possible involvement of state agents and any potential cover-ups. The investigation should not be limited to the conduct of the immediate perpetrator; it should extend, when evidence warrants, to responsibilities arising from orders, omissions in supervision, disregard for known risks, and failures in the organization of services.

VII. Establish external oversight and secure reporting mechanisms

Establish or guarantee an inspection mechanism with functional independence from the custodial authorities, empowered to conduct unannounced visits, access all relevant areas and records, and confidentially interview individuals deprived of their liberty. Its findings and recommendations must be made public and subject to verifiable follow-up.

Establish accessible and confidential channels for reporting violence, lack of medical care, food deprivation, and threats, with effective protection against retaliation. Facilitate access for competent international mechanisms and pay particular attention to facilities where deaths occur repeatedly or where multiple risk factors are present.

VIII. Incorporate differentiated protection measures and prevent discrimination

Adapt care, accommodation, and protection measures to the needs of older adults, people with chronic illnesses, disabilities, or mental health conditions. Incorporate a gender perspective in the supervision of women's facilities and in the investigation of women's deaths, taking into account histories of violence and neglect.

Examine potential differences in treatment based on race, political affiliation, or ties to independent civil society. Ensure that these conditions do not restrict access to food, healthcare, reporting mechanisms, or health-related release measures.

IX. Guarantee public information and records that allow for the prevention and clarification of deaths

Publish regular, verifiable, and disaggregated information on deaths in prison custody, including data on the facility, date, known circumstances, and the status of investigations, while preserving the privacy of victims and their families. Distinguish between preliminary causes, institutional accounts, and the conclusions of independent investigations.

Also publish population and prison capacity data to contextualize the records, evaluate detention conditions and guide preventive measures.

X. Guarantee the rights of families and comprehensive reparation

Promptly report deaths and provide families with clear information about the circumstances, medico-legal procedures, and investigations. Guarantee access to the body and relevant documentation, as well as their effective participation in clarifying the facts, without unjustified obstacles or reprisals for reporting.

Ensure effective resources and comprehensive reparations when human rights violations are determined, including compensation, care for the harm suffered, public clarification of the facts, and guarantees of non-repetition. Reparations must be accompanied by measures to correct the institutional practices that contributed to the death.

The implementation of these recommendations must be formalized in a public plan with responsible authorities, resources, timelines, and verifiable indicators. Its purpose is to prevent further deaths and ensure that deaths in prison custody are investigated, responsibilities are determined, and families have access to justice and reparations.